A comprehensive evaluation and training system for human jump index based on multi-modal signal is composed of a motion information collection module, a motion information processing module, a motioninformation evaluation module and a motion training guidance module, which can analyze and train a human lower limb explosive force, speed and other motion indexes. The device is designed based on nine-axis sensor, which overcomes the shortcomings of single data dimension, big error of evaluation result and single evaluation of traditional evaluation device. In the human-computer interaction interface of the information evaluation module, subjects can clearly see the multi-dimensional motion data and analysis results, and store and share them to the evaluation center. In the human-computer interaction interface of the training instruction module, the instructor can also set the movement parameters by himself, display the movement effect through VR technology, and make the best human jump training scheme. The device has the advantages of small volume, low cost of use and maintenance, high reliability, and is suitable for the evaluation and training of human jumping movement indexes.