Wheat (Triticum aestivum L.) is one of the most important crops in the world. Wheat diseases are a major problem in the production of this cereal, among which fungal diseases cause the greatest damage. Fungi of the genus Fusarium are one of the most common wheat pathogens, which can infect various plant organs, but most commonly they infect the spike. The Fusarium Head Blight (FHB) is one of the most destructive fungal diseases of wheat all over the world, which reduces grain yield through increased flower sterility and poor grain filling and negatively affects the quality properties. In addition, Fusarium species produce mycotoxins which can be harmful for human and animal health, of which the most widespread are deoxynivalenol (DON) and zearalenone (ZEN). Agricultural management practices such as crop rotation, soil tillage and fungicide application are only partially effective in preventing this disease. The most effective method against FHB is to create resistant genotypes. Diallel analysis provides information about general and specific combining abilities of individual parents as well as information on the nature of gene effects for tested traits. The aim of this study was to evaluate the general (GCA) and specific (SCA) combining abilities for visual scores of FHB symptoms as well as for DON and ZEN content, to determine heterosis in F1 crosses and to estimate variance components and heritability for these traits, to determine correlations among traits and to estimate the decrease in the value of agronomic traits in the studied genotypes due to artificial infection. Eight homozygous parents, differing in their resistance to FHB, and their 28 F1 crosses, produced according to the diallel mating design without reciprocals, were included in the study. Field trials were carried out in the breeding nursery of the Bc Institute for Breeding and Production of Field Crops Zagreb at the location Botinec, Croatia during three consecutive years. The trials were conducted each year under artificial inoculation with the fungus F. graminearum and under natural infection (control). For assessing visual symptoms of FHB under artificial and natural infection the percentage of infected spikelets (visual rating index, VRI) and percentage of Fusarium damaged kernels (FDK) were used. In addition, morphological and agronomic traits such as number of days to flowering, plant height, grain yield per plot, 1000 kernel weight, number of grains per spike, grain weight per spike as well as the protein content of the grain were determined. Quantitative analysis of the mycotoxin content by the HPLC-MS/MS multi-analytical method in the infected wheat grain was carried out on samples taken from artificially inoculated trials. Diallel analysis using Griffing approach Method 2 was performed to estimate general (GCA) and specific (SCA) combining ability effects as well as variance components and heritability of visual symptoms of FHB as well as DON and ZEN content. For all analyzed traits, mid-parent heterosis (MPH) as well as better parent heterosis (BPH) were determined. The correlation among traits under both natural and artificial infections was estimated by computing phenotypic and genetic correlations. Relative efficiency of selection conducted under artificial vs. natural infection to improve resistance to FHB was calculated on the basis of estimated heritability of resistance ratings and genetic correlations between the resistance ratings under artificial and natural infection. Analysis of variance for artificial inoculation showed significant effect of genotype for VRI, FDK, DON content, and ZEN content, significant year effect for FDK, DON content and ZEN content and significant genotype × year interaction for FDK. Diallel analysis revealed significant GCA and SCA effects for all traits associated with the resistance of genotypes to FHB under artificial infection, where as under natural infection significant effect was found only for GCA. Significant GCA points to the importance of the additive gene effects in the control of FHB resistance, while significant SCA indicates the existence of a dominant or epistatic gene effects in some cross-combinations. The additive variance component in the present study was in both types of infections for all traits substantially higher than the dominant variance component. Therefore, it can be expected that the choice of resistant parents can provide an efficient response to selection. High negative GCA values for FHB-resistance associated traits in individual parental lines make them desirable parents in future cross-combinations within the breeding programs aimed at improving FHB resistance. The results of the present study revealed negative MPH for VRI in artificial inoculation in 22 and negative BPH in 6 out of 28 crosses. For FDK, negative MPH was found in all 28 crosses and negative BPH in 15 crosses. Negative MPH for DON content was determined in 26 and negative BPH in 6 crosses while negative MPH for ZEN content was determined in 26 and negative BPH in 10 crosses. Negative heterotic effects, determined both in relation to mid-parent and better parent, provide a perspective for classical wheat breeding by creating resistant line varieties, especiaslly in hybrid wheat breeding. Under artificial inoculation high positive correlations were observed among all traits associated with resistance to FHB (VRI, FDK, DON content and ZEN content). Therefore, by monitoring the visual symptoms of the disease in the spike, as fast and inexpensive method, it would be possible to predict the degree of grain contamination as well as the content of mycotoxins. High negative correlations under artificial inoculation were observed between the visual FHB simptoms and both grain yield and its components, such as grain number per spike, 1000 kernel weight, grain weight per spike and test weight. The present study also revealed a high negative correlations between plant height and traits associated with FHB resistance, indicating that taller genotypes are more resistant to the disease. High negative correlations between the traits related to visual symptoms of FHB and grain yield and its components indicated the importance of genotypic resistance to FHB to maintain the yield stability in environments favorable for the development of FHB. In the present study a reduction of grain yield and its components was observed under artificial inoculation relative to the natural infection. As an average over 36 genotypes a relative reduction of 27,2%, 10,9%, 22,4%, 10,5% and 13,5% was observed for grain yield, 1000 kernel weight, grain weight per spike, test weight and the number of grains per spike, respectively. The estimates of heritability for visual FHB symptoms (VRI and FDK), as an important component affecting the relative efficiency of indirect vs. direct selection, were in all three experimental years considerably higher under artificial inoculation than under natural infection. The efficiency of indirect selection under artificial inoculation was higher than efficiency of direct selection under natural infection in all three experimental years for VRI but only in one year for FDK as the consequence of differences in natural disease incidences among years.