With the rapid development of technology, the massive amount of data accumulated in the field of economic management contains rich information, but traditional analysis methods are difficult to effectively explore its complex features and patterns. This study focuses on using CNN models to solve the problem of feature extraction and modeling in economic management big data, and explores its application value and contribution in this field. The CNN model, with its powerful ability to automatically extract features and capture local correlations in data, can break through the bottleneck of traditional methods in processing high-dimensional and nonlinear economic data. By preprocessing economic big data and applying CNN models, key features from multi-source heterogeneous data such as tax data and market transaction data can be efficiently extracted, avoiding the subjectivity and limitations of manual feature extraction. Through statistical analysis, these characteristics not only reveal the significant differences between China’s population and social economy in the eastern and western regions, but also show the overall trend of the economic development of cities and regions through time and space analysis. The application of CNN models provides a new technological path for economic data analysis, which helps to improve the accuracy of economic forecasting, optimize economic management decisions, and is of great significance for promoting the development of research and practice in the economic field.