With the increasing demand for power loads, accurate forecasting of power loads is of great significance for power system planning and generation operation scheduling. Improving the accuracy of short-term power load forecasting can significantly enhance the efficiency of power generation and dispatching, ensuring the efficient operation of the power system. To this end, this paper proposes a short-term power load forecasting model based on Sand Cat Swarm Optimization (SCSO) and Temporal Convolutional Network (TCN). To address the issue of poor initial population quality in the SCSO algorithm, this paper improves SCSO and uses it to optimize the hyperparameters of TCN, thereby significantly enhancing the performance of the neural network by finding the best combination of hyperparameters. The experimental results show that ISCSO-TCN achieves 11.72, 14.21, and 98.49% on MAPE, RMSE, and R2, respectively.