This study aimed to develop and evaluate a deep learning-based Physical Education, Sports, and Health learning module to improve football learning outcomes among fifth-grade elementary school students at SDN 22 V Koto Kampung Dalam. This research employed a Research and Development (R&D) design using the Borg and Gall development model, including stages of needs analysis, data collection, product design, expert validation, product revision, product testing, and final revision. The developed module was validated by material and language experts and tested using pre-test and post-test assessments in experimental and control groups. The findings showed that the module achieved high validity scores, including 88% from the material expert and 83% from the language expert, categorized as highly feasible. Teacher responses reached 91%, indicating excellent practicality. The effectiveness test demonstrated a significant improvement in students’ football learning outcomes, with tcount > ttable (4.06 > 1.833). Therefore, the developed deep learning-based module was proven valid, practical, and effective for improving elementary students’ football learning outcomes.