The gender gap in education in favour of girls is a widely known phenomenon. Boys generally have higher dropout rates, obtain lower grades, and show lower engagement. Insight into factors related to these academic outcomes could help to address the gender gap. This study investigated, for Dutch language classes, (1) how boys and girls differ in behavioural engagement, (2) which teacher support dimensions (autonomy support, structure, involvement) may explain gender differences in engagement (mediation hypothesis), and (3) whether and which of these teacher support dimensions matter more for boys' as opposed to girls' engagement (moderation or differential effects hypothesis). A total of 385 Grade 7 students and their 15 language teachers participated in this study. Teacher support was assessed through student reports. Student engagement was measured using student, teacher, and observer reports. By means of structural equation modelling, the mediating role of the teacher support dimensions for gender differences in behavioural engagement was tested. The potential differential role of the teacher support dimensions for boys' and girls' engagement was investigated through multigroup analysis. Boys were less engaged than girls and reported lower support from their teacher. Autonomy support and involvement partially mediated the relationship between gender and behavioural engagement. Autonomy support was demonstrated to be a protective factor for boys' engagement but not for girls'. Structure and involvement contributed equally to engagement for both sexes. Although involvement and autonomy support partly explained the gender gap in engagement (mediation hypothesis), more support was found for differential effects of autonomy support on boys' versus girls' engagement (differential effects hypothesis).