GREEN TAX PREDICTION AND SUSTAINABLE DEVELOPMENT OF SMALL AND MEDIUM-SIZED COAL ENTERPRISES

Against the backdrop of intensifying global environmental issues, the coal industry, known for its high pollution and energy consumption, is facing increasingly stringent environmental regulations and green tax policies. While large coal enterprises have strong technical and financial support and can relatively easily cope with these pressures, the majority of the industry, which consists of small and medium-sized coal enterprises, faces numerous difficulties in technological innovation and environmental investment. Small and medium-sized enterprises, due to limited resources, often struggle to reconcile the contradiction between economic benefits and environmental responsibility, making the effective response to green tax policies an urgent issue to address. In light of this, this study combines the Autoregressive Integrated Moving Average with Exogenous Variables (ARIMAX) model, the Back Propagation Neural Network (BPNN) algorithm and rolling forecasting technology to design an intelligent green tax revenue forecasting model specifically for small and medium-sized coal enterprises in China. The model is trained and validated using public datasets, and the results indicate that it can effectively predict long-term trends and structural changes while maintaining high forecast accuracy and robustness. The dynamic R-squared values exceed 0.85 in most cases and the model has good performance in rolling forecast accuracy. This study aims to provide more accurate green tax revenue forecasting support for small and medium-sized coal enterprises, assisting them in achieving sustainable development in a complex policy environment. Thus, the study can offer a basis for decision-making by enterprise management and government departments.