The Neretva river valley with its climatic and geographical conditions, is a unique area due to the almost monoculture plantations of Satsuma mandarin (Citrus unshiu Marc.). Due to insufficient pasture research and quality of honey from Satsuma mandarin, objective of this study was to determine the foraging activity of honey bee colonies and the composition of nectar and honey. The three-year research was conducted from 2014 to 2016 in the Neretva valley in the area of Opuzen. The amount of nectar and sugar concentration was monitored on four varieties of mandarins, ‘Kawano Wase‘, ‘Okitsu’, ‘Chahara’ and ‘Zorica’ by sampling 10 open flowers for each variety in a single measurement. Six measurements per day were performed at intervals of two hours (8:00, 10:00, 12:00, 14:00, 16:00 and 18:00). In the final year of research, nectar secretion was monitored on flowers in the opening phase and in the phase of a fully open flower. Nectar sampling was carried out using glass capillaries with a capacity of 75 µL according to Corbet (2003). Based on the collected data, the honey potential of investigated varieties was determined. Monitoring of honeybee visits to mandarin was conducted in April 2016, according to Abrol (2007). Flower visit is monitored by counting bees in the studied varieties. Two trees were selected in each variety. On each of the trees, three branches were marked with the uniform number of flowers. Visits are monitored by counting bees three times a day, at 10:00, 14:00 and 18:00 for 5 minutes per each branch. Floristic observations were performed at selected locations and plant species listed that bloom at the same time as the Satsuma mandarin. Foraging activity is determined using honey bee colonies of Carniolan bees (Apis mellifera carnica, Pollmann, 1879). Returning bees were collected on the hive entrance for the content of honey sac and pollen loads study. The analysis was performed according to the method of Soengen and Jay (1974). Sampling of pollen loads was conducted using outdoor pollen collector set on the hive entrance. From the physico-chemical analysis of the collected samples of honey, the water content, electrical conductivity, pH and activity of diastase, as well as qualitative melisopalynological analysis were conducted. Analysis of volatile compounds was performed using microextraction of headspace on a solid phase (HS-SPME) and ultrasonic extraction with organic solvents (USE) on samples of nectar, honey sac contents and honey. Mean values of the nectar quantity were larger and not significantly different in cultivars ‘Okitsu’ (23.40 mL) and ‘Kawano Wase‘ (22.79 mL), compared to the other two studied varieties. Also, a statistically significant difference was found between the other two varieties, where the average higher amount of nectar is found in variety ‘Chahara’ (21.60 mL) compared to ‘Zorica’ (20.01 mL). The range of the specified parameter in all the years ranged from minimally 5 to maximally 75 µL. The lowest concentration of sugar (18.43%) was found in variety ‘Zorica’ that was significantly different from the nectar of other studied varieties. The mean values of the studied parameter in ‘Okitsu’ (20.35%), ‘Kawano Wase‘ (20.66%) and ‘Chahara’ (20.17%) were not significantly different. The mean values of the quantity of secreted nectar in closed flowers of all varieties ranged from 18.96 to 21.03 µL in open flowers, and are significantly different. Based on the mean values of the sugar concentration, higher concentration was determined in open (21.62%) compared to the closed flowers (19.04%), taking into account all the studied varieties. The honey potential of investigated varieties ranged from 301 kg to 361.5 kg of honey per hectare. There is no exact data on cultivated areas according to different varieties in the Neretva valley. The estimation that different Satsuma mandarin varieties cover 1 700 hectares suggests that it was possible to produce from 511.7 tons to 614.5 tons of mandarin honey in the study period. The highest mean number of bees visits was determined on variety ‘Kawano Wase‘ (12.89), followed by ‘Okitsu’ (11.56) and ‘Chahara’ (11.11), while the lowest (10.00) visit was on the variety of ‘Zorica’. Floristic research has identified a total of 78 taxa (species, subspecies, and varieties). Based on the data on the flowering source, from the determined plants, nine of them were the source of nectar only, 45 of nectar and pollen, while 24 were visited by bees exclusively for pollen. Foraging activity of bee colonies was determined in the pasture based on the analysis of foragers at the hive entrance. In total, 10,470 foraging bees were processed, of which 2595 came with the load to the hive. Nectar was found in 1079, pollen in 1476, however nectar and pollen loads together were found only in 40 bees. 7875 foraging bees returned to hive without load. The highest number of foraging bees with nectar (404) and pollen load (622) was determined at 8:00, while the number of bees at 13:00 and 18:00 was significantly lower. The average weight of the honey sac ranged from 20.22 mg at 8:00 to 21.88 at 13:00, and the range was from 0.70 mg to 58.60 mg. However, comparing the resulting average values, there were no statistically significant differences among investigated wights per times (hours). Average maximum weight of pollen loads was found at 8:00 am (8.47 mg) and was not significantly different from the values measured at 13:00 (8.44 mg). The mean weights of pollen loads collected at 18:00 (7.67 mg) were significantly lower in comparison to two previous samplings. The range of pollen loads weight ranged from 0.50 mg to 45.90 mg. The mean total weight of pollen collected in external pollen collectors during 2014 was 30.92 g in the morning and 8.78 g in the afternoon measurements, and was significantly different. In 2015, the average weight of collected pollen in the morning was 10.55 g and was not significantly different from pollen collected in the afternoon (5.18 g). In 2016, the average weight was higher at the site of the Buk-Vlaka (19.79 g) in relation to the location Ušće (12.89 g). According to the botanical origin, collected pollen originated from 22 plant species, among which the most common were orange (Citrus sinensis), rough hawksbeard (Crepis biennis) and evergreen oak (Quercus ilex). In the pollen spectrum of honey sac content, pollen of Satsuma mandarin was determined in 21% of samples, while 79% of the samples did not contain pollen from Satsuma mandarin. Apart from the mandarin, pollen from additional 17 plants was found. In pollen spectrum of honey samples besides Satsuma mandarin pollen the most common pollen of nectariferous plants was from shepards purse (Capsella bursa pastoris), plants of the Asteraceae family (Asteraceae Taraxacum form), evergreen buckthorn (Rhamnus alaternus) and orange (C. sinensis), and of non nectariferous plants pollen of manna ash (Fraxinus ornus), holm oak (Q. ilex), oak (Quercus spp.), orchard grass (Dactylis glomerata) and olive (Olea europea). The average values of the water content in the samples of Satsuma mandarin honey ranged from 17.1% to 18.1%, the electrical conductivity from 0.20 mS/cm to 0.27 mS/cm, pH values from 3.82 to 4.06 and diastase activity from 7.78 to 12.02 DN. The main identified compounds in volatile peaks in nectar headspace were: linalool, αterpineol, 1H-indole, methyl anthranilate and phenylacetonitrile. Relevant extracts contained 1H-indole, methyl anthranilate, 1,3-dihydro-2H-indol-2-on, and caffeine. The main headspace compounds in the honey-sac were: linalool, α-terpineol, 1,8-cineole, 1Hindole, methyl anthranilate and cis-jasmon. The specific compounds of the honey sac extract (USE) were: caffeine, 1H-indole, 1,3-dihydro-2H-indol-2-on, methyl-anthranilate and phenylacetonitrile. However, the composition of the headspace of honey was significantly different in comparison to the representation of phenylacetaldehyde and linalool derivatives that have arisen as a result of conditions in the hive and honey bee activity which is an original contribution to the knowledge of the transport of volatile compounds from the nectar over honey sac to honey. Extracts (USE) of honey contained higher aliphatic compounds as the main component, that are not useful for determining the botanical origin of honey as may be derived from the hive. Traceability of Satsuma mandarin honey is determined through chemical markers: phenylacetaldehyde, phenylacetonitrile, linalool and linalool derivatives, 1H-indole, 1,3-dihydro-2H-indol-2-on, and caffeine. 1H-indole, 1,3-dihydro-2H-indol-2-on and phenylacetonitrile that were found as typical compounds of Satsuma mandarin honey, and not found in other types of Citrus (Citrus spp.) honey.