Aim. The purpose of the study is to predict the occurrence or reveal changes in the electrocardiogram (ECG) of swimmers with the help of stabilometric indicators. Materials and methods. Sixty qualified athletes participated in the study (CMS, MS, age 16–18 years). Bicycle ergometry was used to detect cardiac rhythm and conduction abnormalities, and a stabilometric complex was used to assess postural balance. Results. Incomplete right bundle branch block (RBBB) was observed in 53.33% of athletes; first-degree sinoatrial (SA) block – in 10.01%; extrasystoles (supraventricular: atrial, atrioventricular) – in 33.33%; pacemaker migrations – in 3.33%. Athletes in group 0 recorded a higher level of fluctuations in the center of pressure in the frontal plane in all tests (p ≤ 0.05): COP-RMSD in FP MS EO (> 132.32%), LHT (> 31.94%), RHT (> 76.82%), MS EC (> 83.56%), LHT EC (> 68.61%), RHT EC (> by 87.65%). Using Random Forest machine learning algorithm, it is possible to predict the occurrence or detect changes in the heart rhythm and conductivity in swimmers. Conclusion. The features of changes in the electrocardiogram of athletes were considered, the compared parameters of stabilometry data with and without ECG changes were significant at p < 0.05. A Random Forest model was created to predict or detect changes in the ECG of athletes with the help of stabilometric indicators.