<p>Against the backdrop of global sustainable development, e-commerce has emerged as a pivotal driver of economic growth, and its implications for green productivity merit in-depth scholarly investigation. Using panel data from 252 prefecture-level cities in China between 2012 and 2022, this study constructs a comprehensive indicator system via the Entropy Weight-TOPSIS method and employs a two-way fixed effects model alongside a mediation effect model to systematically examine the impact of e-commerce on urban green productivity. The empirical findings are as follows: First, e-commerce development exerts a statistically significant positive effect on urban green productivity, with an estimated coefficient of 0.397. Second, digital technology innovation serves as a significant mediating variable between e-commerce development and urban green productivity, accounting for 4.3% of the total indirect effect. Third, this influence exhibits notable regional heterogeneity, displaying a pattern of “strongest in Central China (coefficient = 1.550), followed by Western China (coefficient = 1.073), and weakest in Eastern China (coefficient = 0.281)”. A series of robustness checks and endogeneity tests further support the reliability of these core results. Moreover, this study innovatively incorporates a random forest machine learning model into the analytical framework; the predictions generated by this model align with those from the benchmark regression. Overall, the results suggest that China’s urban green productivity is likely to enter a phase of rapid improvement during the 2022–2025 period. These findings offer substantial policy and practical implications for enhancing China’s green productivity and promoting broader societal sustainable development.</p>