The promotion of regular exercise and the adoption of lifelong activity habits is undoubtedly one of the primary goals of school Physical Education (PE). To this direction, game-based PE teaching approaches are utilized by practitioners as a means for both launching students’ in-class Physical Activity (PA) levels and promoting wider-base curriculum learning outcomes. As an approach of this kind, the Teaching Games for Understanding (TGfU) model is widely used as a game-based approach aiming at developing students’ tactical skills and knowledge. Up to nowadays, only a limited amount of research has tested the impact of TGfU on students’ PA levels during PE instruction time. Therefore, the aim of the present research was to evaluate the effects of a TGfU intervention program on primary school students’ patterns of PA. Sixteen third and fourth grade students (mean age=9.2 years, SD 0.7) from two public schools in Athens, Greece, participated in a TGfU intervention program, that took place during their regularly planned PE classes. Students’ PA levels were assessed at baseline and six-weeks follow-up, within a 10-minute four vs four handball game. PA measures were obtained both quantitatively by using pedometers (steps per minute) and qualitatively through direct observation with the System for Observing Fitness Instruction Time (SOFIT). The comparison of students’ pre-and-post measures showed that although students’ movement intensity was higher at the end of the intervention, their PA behaviors did not differ significantly between the two time points. These findings were attributed both to the development of students’ game skills and competence and the short duration of the evaluation period. Results suggest that judgments about the health-related usefulness of the TGfU model should be based on multiple-source and long-duration assessment procedures.