Aiming at the uncertainty of random factors of load level in the power system, a probability analysis method for small disturbance stability of the power system based on the unscented transformation method is proposed. The method transforms the probability analysis problem into a deterministic analysis problem through a symmetric sampling strategy, and then calculates the probability index of small disturbance stability according to the weight distribution of each sampling point by the unscented transformation method. The example analysis is verified based on the power system analysis toolbox (PSAT) 3-machine 9-bus system in MATLAB. Then, the probability analysis of small disturbance stability for a 10-machine 39-bus system is performed. The validity and practicability of the proposed method are verified by comparing the unscented transformation method with the two-point estimation method. The simulation results show that the relative errors among the unscented transformation, Monte Carlo and two-point estimation methods are small, but the unscented transformation method takes much less time than the Monte Carlo method, which enables the rapid evaluation of small disturbance stability probability analysis and meets the requirements of engineering applications.