A reliability assessment with incomplete performance degradation data under small and non-failure samples was made. A data-based random number simulation method proposed by Taylor and Thompson was used to complete the missing sample degradation data

the Bootstrap method was used to expand the sample size

and the reliability assessment of the bearings was analyzed based on the degradation trajectory-based method. The vibration degradation of 7 groups of tested bearings was selected and analyzed under the conditions of complete and incomplete performance degradation data

respectively. It was found that the method proposed can obtain accurate reliability assessment and the absolute value difference between two results was controlled within 01. The above results were compared with the results obtained by analyzing the lifetime data through the weighted E-Bayes method and the maximum likelihood estimation. It was found that the results obtained by the method were better

and the actual error was within 10% compared with the actual test

showing positive practical significance for improving the evaluation accuracy and reducing the test cost.