Peer victimization and anxiety are consistently positively correlated, though the longitudinal relationship remains inconsistent. Previous research often failed to account for the reciprocal relationship between traditional and cyber victimization, with limited evidence of the influence of broader societal factors. Thus, the present study aimed to longitudinally examine the relationship between victimization, cyber victimization, and anxiety within the specific context of the COVID-19 pandemic during one school year (2021/22), employing both the cross-lagged panel model and the random-intercept cross-lagged panel model. The study included 1766 students (58.5% females; M[subscript age]= 15.33; SD = 1.20) from lower- and upper-secondary schools in Slovenia who participated in three time-points with approximately 8 weeks apart, with T1 occurring before school closures, T2 during school closures (which lasted 17-21 weeks), and T3 after school closures. Findings revealed that the relationship between anxiety and victimization changed across the school year. The results of the cross-lagged panel model indicated that traditional victimization in T1 positively predicted anxiety in T2, and anxiety in T2 positively predicted traditional victimization in T3. Cyber victimization at T2 predicted lower anxiety at T3. The results of the random-intercept cross-lagged panel model showed that t the within-person level, students experiencing more cyber victimization than expected reported greater subsequent victimization, especially after school closures. The study provides insights into the dynamics of the relationship between peer victimization and anxiety during a specific period of school closure. The findings do not fully align with any single theoretical model, highlighting the complexity of these interactions in the pandemic context. This study underscores the need to consider broader socio-environmental factors in understanding the relationship between peer relationships and psychosocial outcomes.