Hydraulic pumps belong to high reliability and long life product, its reliability has a critical influence on the reliability and performance of whole machine. After a long accelerated life test, it may still be nonfailure for high reliability and long life products, which makes it difficult to assess reliability. Aiming at the limitation of the traditional reliability assessment techniques, the reliability of the hydraulic pumps shorttime test method was proposed, which based on performance degradation theory to predict the reliability and residual life of products. To establish residual life prediction model for hydraulic pumps, based on the volumetric efficiency data, the nonlinear leastsquare method was adopted to fit time series, simultaneously, the optimum path search algorithm was built to select the best degradation path. The consistency check of the data in prediction model and the stepstress reliability life test were done by using nonparametric statistical theory, and the overall sample life distribution function was constructed with Monte Carlo numerical simulation method. The results show that the simulation data of fault sample distribution are consistent with stepstress reliability life values. Comparing the test parameters curves of the shortterm test with stepstress accelerated life test, the correctness and feasibility of the shorttime test method are further verified.