It is of great significance to consider uncertainty factors and correlation factors in order to assess load flow characteristics of power system accurately and comprehensively. Copula theory is introduced to establish the probability distribution model of correlated input random variables. A method to obtain input random variable’s empirical cumulative distribution function and inverse function based on measured discrete data is proposed to handle the input random variable whose marginal distribution doesn’t follow common distribution function. Copula theory is applied into Monte Carlo simulation method, and a probabilistic load flow (PLF) method which can deal with the correlation between input random variables flexibly is proposed. Taking wind power for example, the accuracy of probability distribution model of correlated input random variables established by Copula theory is evaluated. The validity of proposed PLF method is tested on the IEEE 57 bus system with wind power. The simulation results show that the proposed method is effective and accurate.