With urban development and population growth, building load management has become increasingly important. Accurate prediction of building load can be used to develop strategies to reduce energy consumption and improve energy efficiency. In this paper, Python was used as the programming language, and Long Short-Term Memory (LSTM) neural network was used to realize the prediction of building electricity load. Firstly, the building electricity load data were collected and preprocessed, including data cleaning, normalization and other operations. Then, the LSTM neural network was trained on the historical electricity load data in order to accurately predict the future load situation. Finally, the LSTM model was verified to have a good effect in the task of building electricity load forecasting.