There are always some different correlations between wind farms in a power grid.This paper uses inverse Nataf transformation to generate correlated wind speed samples to get the correlating wind farmsgenerations.Considering the uncertainty of wind speed，the actual production of each wind power plant is considered as a negative Weibull random distribution load and characterized by its historical time series data.The spatial correlations of wind power plants are properly modeled through stationary variance-covariance matrices.Wind farms are included in modified IEEE30-bus and IEEE118-bus systems.Monte Carlo simulation is adopted to quantitatively analyze how the indicators of optimal power flow results volatilities as a result of an increasing correlation among wind farms.