Under the background of high launch density, high interpretation requirements and high amount of data of launch vehicles, the traditional automatic interpretation criteria with incomplete coverage, high threshold of criterion design and time-consuming execution, which are increasingly prominent, and the lack of general interpretation algorithms to supplement affect the effect evaluation and system performance evaluation of launch vehicles. In order to fully mine the parameter variation law implied in the massive telemetry data of launch vehicles, the intelligent interpretation algorithm is designed as a supplement to the traditional algorithm, and improve the interpretation coverage and execution efficiency of the traditional interpretation. Taking the telemetry data generated by the long-term power test of liquid launch vehicle as the research object, the integrated neural network intelligent interpretation algorithm is designed. Under the given interpretation index, it is concluded that the integrated neural network is suitable for abnormal frequency and frame loss interpretation scenario where traditional criteria are difficult to interpret. The interpretation performance is improved by 30%, and the coverage of existing criteria are improved. Later, it could provide research examples for the improvement of interpretation system and the study of intelligent interpretation application.