The transmission congestion probability forecasting for power systems with high level wind power integration is researched， and an optimal boundary power flow and a probabilistic dynamic forecasting method for ultra-short term transmission congestion are developed. Probability distribution is applied to express the uncertainty due to wind power integration. The proposed forecasting method applies the optimal power flow algorithm as the section power flow calculation framework， and obtains the boundary and probability distribution of section power flow through boundary load flow and Monte Carlo simulation in which antithetic variable sampling is used to reduce sampling frequency， and then the congestion probability of transmission sections is predicted. Simulation results on IEEE 39-bus and IEEE 118-bus systems validate the efficiency of the proposed probabilistic forecasting method.