<p>The sports physical activity sector is continuously searching for innovative solutions to traditional issues like maintaining fitness through ideal posture in a correct form to enhance training programs and boost fitness performance. The effectiveness of training programs depends significantly on achieving and maintaining correct exercise posture, which is challenging without getting appropriate feedback. However, many college students struggle to obtain ideal posture without guidance from professionals, which may be quite expensive. The fuzzy apriori miner for physical fitness recognition (FAM-PFR) algorithm is introduced to recognize and provide real-time feedback on student training postures using association mining-based video recognition with an improved fuzzy Apriori algorithm. The FAM-PFR algorithm utilizes fuzzy logic to deal with ambiguities in posture data derived from video frames, such as correct posture forms with membership degrees. Distances, landmarks, and angles create fuzzy Apriori item sets and association rules. The fuzzy Apriori concept is refined by pruning to maximize processing performance. The system gives tailored training advice and real-time exercise form feedback on different frames. Compared to Fuzzy Apriori and Decision Trees, Apriori, and Fuzzy Mamdani Decision System, FAM-PFR has faster execution time, better recognition accuracy, personalized recommendations, and fuzzy rule generation efficiency. The system recognizes physical activity and provides fitness association rules. Data-driven, individualized fitness recommendations in collegiate fitness programs could result.</p>