Large scale renewable power generation brings great challenges to the power system operation and stabilization. Energy storage is one of the most important technologies to face the challenges. This paper proposes a method for transient stability probability evaluation of power system with wind farm and SMES. Firstly, a modified 11-bus test system with both wind farm and SMES has been implemented. The wind farm is represented as a doubly fed induction generator (DFIG). Then a stochastic-based approach to evaluate the probabilistic transient stability index of the power system is presented. Uncertain factors include both sequence of disturbance in power grid and stochastic generation of the wind farm. The spectrums of disturbance in the grid as the fault type, the fault location, the fault clearing time and the automatic reclosing process with their probabilities of occurrence are used to calculate the probability indices. With the proposed method based on Monte-Carlo simulation and bisection method, system stability is "measured". Quantitative relationship of penetration level, SMES coil size and system stability is established. Considering the stability versus coil size to be the production curve, together with the cost function, the coil size is optimized economically.