The purpose of this research is to analyse the influence of enviromental disclosure and enviromental performance on economic performance with firm size as control variable. Enviromental disclosure was measured by GRI index, enviromental performance measured by PROPER index, firm size measured by Ln total assets and economic performance measured by economic performance index. The sample of this study consists of 32 companies listed on the IDX 2013 - 2016. The criteria of the research sample are (1) Manufacturing companies that follow PROPER, (2) Issuing financial statements with rupiah currency, (3) Publish a complete annual report. The results of this study inform that the performance of Multivariate Adaptive Regression Spline (MARS) is better than multiple linear regression. The result of multiple linear regression informs that not all classical assumption requirements are met. This results in a non-significant regression model, small R-square and many predictor variables have no effect on response variables. MARS is one of the alternative methods to overcome the lack of multiple linear regression method. MARS is not a requirement with classical assumptions because it includes one of the non-parametric regressions. MARS results informed that the MARS model is significant, R-square is large and the variables that affecting the Economic performance are enviromental disclosure and enviromental performance while the most influential is the enviromental performance.