Abstract: In order to avoid the risk of power flow thermal over-limit caused by unreasonable capacity expansion of transmission lines, a day-ahead dispatching model of power system considering transmission line dynamic capacity expansion risk is proposed. The model aims to reasonably improve the consumption capacity of various resources under the guarantee of safe power transmission. Firstly, the dynamic thermal rating (DTR) prediction model of transmission lines is established based on the ForecastNet neural network with time-variant structure. The prediction model can dynamically track the influence degrees of environmental factors and correct the prediction network weights to improve the prediction accuracy. Secondly, based on the DTR day-ahead forecasting results, the dynamic capacity increase margins of transmission lines are determined, and the day-ahead scheduling model for electricity market bidding mechanism is constructed by introducing the capacity increase risk cost and the risk preference coefficient. Finally, the proposed model is simulated on the IEEE-39 nodes system by using the real network data of Liaoning Province. The simulation results verify the accuracy of the prediction model and the effectiveness of the scheduling model. At the same time, the simulation results show that the scheduling model can make full use of the redundant transmission space of the transmission network, greatly enhance the consumption level of renewable energy, and ensure the security and economy of the day-ahead scheduling strategy in the electricity market environment.