By using wind speed probability distribution，wind disturbance histogram and correlation parameter as constraints，a wind modeling method considering correlation and statistical characteristics is proposed.Considering monthly10-min averaged wind speeds usually follow a mixed Weibull distribution，the single month wind speed probability distribution is firstly divided into several single Weibull distribution with different number of samples.Then，several time series with different number of samples，special correlation parameters and single Weibull distribution characteristics are generated.Final time series can be obtained by combining generated time-series and adjusting the orders of samples.Simulation results show that the simulated wind speeds can both reflect its own and nearby wind farms statistical laws.