Stochastic finite element method and reliability technique are used to determine the safety degree of the turbine blade with parametric uncertainty. The material, geometric parameters and rotating speed of blade exhibit notable random fluctuations, so the conventional deterministic analysis of blade can’t provide complete information. The stochastic analysis can tackle the uncertainties in structural parameters and obtain the probabilistic characteristic of the vibration characteristic. In this paper, the study focuses on the reliability assessment of the blade with uncertainty parameters based on the stochastic finite element method (SFEM) and the mean-variance method. The perturbation stochastic finite element method (PSFEM) is used to calculate probabilistic characteristic of the natural vibration of the turbine blade. Based on the stochastic finite element method, the mean-variance method is used to calculate the resonance reliability of the blade. The example shows that the present method is valid.