Soil erosion is globally primary cause of soil degradation, which is directly responsible for long-term reductions in agricultural productivity and significant financial losses. Due to the significant spatial distribution and diversity of erosion forms, the presence of harsh climatic predisposing factors and the specific bare karst landscape, Pag Island is ideal for a systematic study of the occurrence and intensity of soil erosion. The main goal of this doctoral dissertation is development of a multi-resolution methodological approach for the systematic study of soil erosion processes, based on the use of multiple research levels, with different spatial coverage and research objectives. The methodological framework of the developed multi-resolution approach for the study of soil erosion enabled the study of different aspects of the soil erosion process, through three different research levels. The macro level of research includes the process of creating a high-resolution multi-criteria GIS soil erosion susceptibility model, for the entire area of the Pag Island. Monitoring of multi-year (2014. – 2018.) and one-year (2019. – 2020.) soil erosion intensity based on VHR models was conducted within the meso level of research. The micro level of research includes activities related to the optimization of interval TLS measurements and detection of subcentimeter spatio-temporal changes caused by uphill headcut retreat. Implementation of the developed multiresolution approach for the study of soil erosion resulted with the achievment of all scientific goals set within this doctoral dissertation. The first high-resolution (1 m) GIS-MCDA model of soil erosion susceptibility was created, on the basis of which the most endangered parts of the island were isolated. Ultimately, it was found that about 40% of the surface of the Pag islands belongs to the high or very high susceptibility areas. Active traces of soil erosion were determined by monitoring the long-term intensity of soil erosion in three of the five gullies analyzed within the Santiš peninsula. Great variability in spatial coverage and total amount of eroded material was found. Most spatio-temporal changes (about 85%) were detected within the two headcuts of the gully Santiš. The determined oneyear volumetric intensity of soil erosion within the Santiš ravine was 13.46 m3. The correlation between the determined one-year intensity and the mapped lithological units showed that the highest spatio-temporal changes have occurred within the carbonate sandstones and sands and brown soil deposits. There was a pronounced annual variability in the intensity of soil erosion, which was correlated with the oscillations in the amount and distribution of precipitation. Within the micro level of the research, optimization of interval TLS measurements and UAV photogrametric surveys was performed. While the optimization of TLS measurements enabled a very high (> 90%) coverage of the complex gully morphology, optimization of UAV photogrammetric survey achieved a very high accuracy of created models (RMSE < ± 2 cm). The developed multi-resolution approach for the study of soil erosion was applied as part of this dissertation in a complex karst area, where all scientific research goals were successfully achieved. Apart from similar karst areas, the developed multi-resolution approach is applicable to all other parts of the Mediterranean and the World, affected by the intense soil erosion.