Abstract: Ice storm disaster has various negative impacts on power grid security and stability. To evaluate the power grid risk accurately, a risk assessment method for power grid icing based on short-time icing forecast is proposed. The ice and wind loads on lines can be forecasted by the icing forecast system. The time-varying failure rates of transmission lines can be calculated by the timevarying structure reliability theory. A restoration time model of transmission line with DC deicer is built as is a time interval model of deicer. Owing to the incapability of reflecting the time-varying failure rate of iced lines using the traditional Monte Carlo method, a sequential Monte Carlo method of solving this rate is proposed. The proposed model is tested on an IEEE RTS-79 system. The effects of the time interval of icing forecast and deicer on risk indices are analyzed and related suggestions made.