The trend of market house prices is influenced by various factors, and house price prediction algorithm remains a very classic and challenging nonlinear problem in data analysis. Analyzing various factors that may affect market house prices can help to provide a more accurate assessment of future trends in house prices. Multiple nonlinear regression is suitable for analyzing data affected by many factors. It is more efficient and practical to predict the house price by multiple independent characteristics than only one independent variable. This article is based on a deep learning algorithm and uses the PaddlePaddle experimental platform to learn about the past period of house sale prices and other related data in Beijing. Multiple nonlinear regression methods are used to analyze the data and predict the future house price trend in this region.