There exist correlation and conditional dependence in wind vectors（wind speed and wind direction）between two adjacent wind farms.Simulation of the wind speed and wind direction samples with correlation and conditional dependence in reliability evaluation of a system with a wind farm is the prerequisite for the reliability evaluation.A wind speed correlation model and a wind direction correlation model are built by using copula theory.Based on empirical distribution and Weibull distribution models，the clustering method，transformation method，and interpolation method are used to build the conditional dependence model of wind speed and wind direction.Then，a wind speed and wind direction generation model considering their correlation and conditional dependence is obtained.The proposed model is verified using observed wind data，the results showing that simulated and observed wind data are identical in correlation，conditional dependence，statistics characteristics，and distribution characteristics.In other words，the results have verified the validity and availability of the proposed model.Finally，the proposed model is applied to an IEEE-RBTS test system for reliability analysis of the combined generation system with wind farms.The wake effect is also considered in the reliability evaluation process.The results show that the proposed model can be used to directly evaluate the reliability benefits and analyze the impact of wind farms.