Focusing on the value of green innovation efficiency in cities, this paper explores the intrinsic mechanisms of tax incentives and the innovative development of digital economy industries in promoting efficiency. This paper first proposes the principle of the regression algorithm required for the study as its theoretical foundation. In the principle, it focuses on the solution method of the LS method, proposes the univariate regression model and the multivariate regression model respectively, and determines the parameter estimates through the residual vector derived by the least squares method. Combined with the regression model, the mediation effect model is proposed to examine the impact of tax incentives on urban green innovation efficiency by taking government subsidies and tax incentives as the entry point of tax incentives. The results show that the regression coefficients of government subsidies and tax incentives are 1.769 and 1.838, respectively, which are positive, and the tax incentives can significantly improve urban green innovation efficiency and can realize the improvement of urban green innovation efficiency by promoting the innovative development of the digital economy industry. The effect size of tax incentives on urban green innovation efficiency in different regions of East, Central, and West China shows regional heterogeneity.