The traditional method is affected by massive sports training data, and there is a problem of data capture error. In order to solve this problem, an intelligent capture method of sports training track data based on wearable sensors is proposed. Stratified the continuous interaction space of sports training, decomposed high-level interaction behavior, and obtained the degree of freedom of interaction. Combined with wearable sensor action recognition technology, the original signal is converted into frequency domain signal, and comprehensive indicators are extracted by principal component analysis. Train the index data in the neural network to identify sports training actions. Load balance processing of track data of wearable sensors, combined with backstepping method to control training track tracking error, ensures that intelligent capture of sports training track data will not be affected by tracking position error. Through iterative data processing, the training track data can be captured intelligently. Take football as an example for experimental analysis. From the experimental results, it can be seen that there is a maximum error of 3 m/s between the speed capture result of human motion mechanics of this method and the experimental data. The football movement track is consistent with the experimental data, which has an accurate capture effect.