<p>While extensive research has been conducted on victim-offender overlap, few studies have explored how this overlap manifests across both offline and online contexts.&#xa0;Grounded in General Strain Theory (GST) and Self-Control Theory (SCT), this study analyzes data from 769 rural-to-urban migrant working adolescents aged 16 to 19 from the fourth sweep of International Self-Report Delinquency Survey (ISRD4) in China. Focusing on a socially disadvantaged group often overlooked in recent studies, the study aims to identify distinct groups based on patterns of offline delinquency, problematic online behaviors, and victimization across offline and online domains, and to assess the impact of conventional risk factors. Latent class analysis identified three groups: (1) a low involvement group, (2) a group characterized by higher levels of problematic online behaviors with low victimization, and (3) a high-overlap group across all forms. Only a small proportion (4%) were classified into the high-risk overlap group. Multinomial logistic regression further demonstrated that higher levels of information strain, risk-seeking, and revenge tendency significantly increased the likelihood of membership in the problematic online behavior group or the overlap group, supporting GST and SCT. In contrast, health-related strain reduced the likelihood of problematic online behaviors, suggesting a protective effect. These findings do not support the assumption that migrant working adolescents are a uniformly high-risk group, and show the role of strain and self-control factors in shaping the overlap between problematic behaviors and victimization. They underscore the need for tailored prevention strategies and inclusive policies that incorporate digital well-being initiatives to promote healthier routines, balanced online engagement, and improved support for vulnerable youth in a rapidly digitizing society.</p>