The invention discloses a wind-power grid-connected system risk assessment method based on the wind speed time period feature. The method comprises the following steps: processing the traditional annual wind speed probability distribution as the time-varying distribution function with time-dependent wind speed, reflecting the main change rule of the wind speed by using a time period fitting function, and reflecting the fluctuant feature of the wind speed by using the random variable obeying some probability distribution, wherein the time period fitting function and the wind speed random variable are overlapped to obtain a time period feature model of the wind speed; computing the time-varying capacity of a wind power plant through a function relation between the time-varying wind speed and the wind generation set, and assessing the time-varying risk index of the wind-power grid-connected system by using the Monte Carlo sampling method. By using the method disclosed by the invention, the time-dependent system risk index can be computed, and the auxiliary information can be provided for the power grid planning and designing, the system scheduling and operation-maintenance decision-making; and the method is in favor of managing and controlling the power grid running risk, and improving the reliability level for the power system running.