Background. The importance of monitoring work-related stress and its consequences for mental health is underlined by the increasing prevalence of absence from work due to stress-related illnesses. Objectives. The aim of the study was to examine personal resources as mediating factors in the development of anxiety and depression from work-related stress in high-risk professions. Sample and methods. The sample comprised a total of 208 police officers, prison guards and customs officers (78.8% men, an average age of 36.5) who provided socio-demographic and job-related information. The Occupational Stress Inventory was used (OSI-R), in particular Occupational Roles Questionnaire and Personal Resources Questionnaire, State Trait Anxiety Inventory (STAI X-1), and Zung Self-Rating Depression Scale (SDS). Occupational roles represent measurement of occupational stress. Statistical analysis. Structural modelling was used for statistical analysis. Results. The final model, comprising working time (emergency care, overtime, night shifts), occupational roles and personal resources, predicted 35% of variability of anxiety, and 42% of variability of depression. Significant predictors of anxiety were the roles of overload and ambiguity, which were mediated by recreation and rational-cognitive coping strategies. Significant predictors of depression were the roles of overload, insufficiency and ambiguity, which were mediated by recreation, social support and rational-cognitive coping strategies. Night shifts, overtime and emergency care did not predict anxiety and depression although they predicted occupational roles. The effect of occupational roles including overload, insufficiency and ambiguity on anxiety and depression was statistically significant, yet indirectly, by reducing personal resources. Personal resources are considerable mediating factors between stress, anxiety and depression in the studied professions. Limits. The study was cross-sectional. Causal relationships can be further tested using interventional studies.