 This paper proposes a two-stage short-term load forecasting method called SVLMFE, which combines Sparrow Search Algorithm, SSA, Variational Mode Decomposition, VMD, Long Short-Term Memory Network, LSTM, Multiple Linear Regression, MLR, and Feature Engineering, FE. The first stage of forecasting uses SSA to select the optimal combination of parameters for VMD, while the second stage uses MLR and FE to predict the load value at the moment that has a great influence on the load value at the time to be predicted. The experimental results show that SVLMFE has high accuracy and stability. This article was authored by Wei Jianhuang, Qisong, and Yuanhuang.