China’s economy is in a state of rapid development in today’s digital era, and big data technology has become a key force in promoting industrial upgrading and economic operation. The article first constructs 14 economic operation indexes in 7 aspects to establish the regional economic index system according to each type of regional economy. After researching BP neural networks, a TMBA-BP neural network prediction model based on adaptive t-distribution has been proposed. The prediction model is used to predict the economic development of Liaoning City, and the results show that the TMBABP prediction model is accurate. Finally, the Gini coefficient is combined with the TMBA-BP model to explore the causes of regional economic differentiation in China. In terms of the contribution rate of regional differences, the inter-regional contribution rate is higher than the intra-regional contribution rate. It indicates that in the big data environment, the main reasons for the obvious regional economic differentiation in China are inter-regional differences, followed by intra-regional differences.