Probabilistic small disturbance stability analysis based on Monte Carlo state sampling method is implemented with consideration of uncertainties in generating dispatch， load demand and network topology. Discrete and normal distribution random variables are used respectively to describe the stochastic characteristics of component states and load levels. Numerical characteristics and probability distributions of eigenvalues， eigenvector and participation factors， and instability probabilities are calculated as stability indices. The influence of network topology uncertainties on these indices is discussed as well. Simulation results of a two-area system demonstrate that the proposed method can provide comprehensive stability assessment and robust power system stabilizers design in a statistical sense.