ABSTRACT Homework is a widely used educational practice, but the ways in which students approach it vary considerably and influence its effectiveness. Understanding these behavioral differences is crucial to designing effective and appropriate homework practices. The purpose of this study was to examine students’ homework behavior and determine its relationship with variables such as academic achievement, homework motivation, age, and gender. In the study, homework profiles of students were created on the basis of indicators of homework time, effort, and completion. The study was conducted among 642 middle school students. Data were collected on students’ homework behavior, homework motivation, and academic achievement in mathematics. In the study, four different homework profiles were identified using latent profile analysis: Full Participants, Consistent Completers, Low Effort Contributors, and Minimal Participants. The findings revealed significant differences in academic achievement and motivation across the identified homework profiles. Students in the Full Participants profile (high effort, completion, and time) demonstrated the highest levels of achievement and intrinsic motivation, whereas extrinsic motivation was more pronounced in the Low Effort Contributors profile (low effort, moderate time, low completion). These results suggest that students’ motivation and achievement are closely associated with their homework engagement patterns, particularly the levels of effort and completion. In addition, younger students were more likely to be in high‐engagement profiles, while older students tended to cluster in profiles characterized by lower homework engagement. Gender differences also emerged, with male students more frequently represented in both the highest (Full Participants) and lowest (Minimal Participants) profiles, and female students more commonly found in the Low Effort Contributors profile. The study emphasizes that teachers should consider individual differences in students when assigning homework.