Graduate students in mental health service training programs are at risk of experiencing burnout (Kaeding et al., 2017). However, limited research has examined how insecure attachment affects burnout of graduate trainees. Guided by adult attachment theory, this study examined a conceptual model depicting the relations among attachment insecurity, selected psychological variables, burnout, and psychological distress. Specifically, this two-tier indirect effect model hypothesized that insecure attachment would be associated with lower self-compassion and more emotion regulation difficulty (the first-tier mediator) which in turn, would result in higher burnout (the 2nd-tier mediator) and eventually greater psychological distress. The final sample included 216 U.S. and 201 Taiwanese mental health graduate trainees. Structural equation modeling (SEM) was used to analyze the data. Findings from the final models best supported by the data revealed that in both cultural groups, attachment anxiety contributed to lower self-compassion which subsequently resulted in greater burnout, but that emotion regulation did not play a mediational role in the models. Results also demonstrated cultural differences in several paths of the proposed model. For the U.S. sample, only higher attachment anxiety indirectly contributed to more psychological distress through low self-compassion, and burnout was best positioned as another psychological distress indicator, instead of being a mediator. However, for the Taiwanese sample, both attachment anxiety and attachment avoidance demonstrated significant indirect effects on higher psychological distress through lower self-compassion and burnout was a 2nd-tire mediator through which attachment anxiety indirectly contributed to higher psychological distress. These findings advance our understanding of the role of adult attachment insecurity in the development of burnout and psychological distress for graduate trainees in mental health fields, as well as the possible cultural differences in the observed variables and their relations. Counseling implications, limitations, and future research directions will be discussed.