Abstract: In power system,many random variables are correlated with each other,so ignoring correlation may cause calculation errors,even have a direct impact on safe and economic operation of the power system.However,existing method of theoretical calculation is not accurate enough to describe random variables,difficult to deal with its correlation and consuming computing time.Thus,random variables simulation based on Copula theory was put forward.Based on the concept of rank correlation coefficient,Monte Carlo（MC） simulation method,and Copula theory,it built sample sampling method considering random variables correlation.This method can deal with the correlation between random variables of normal and non-normal distribution.Then,the median Latin hypercube sampling techniques were put into the method proposed which raised calculation efficiency,and improved accuracy and robustness of the method.Finally,we can take the evaluation of available transfer capability（ATC） as an example,and used risk concept to verify the effectiveness of the method by improved IEEE 30 test system.