Dynamic analysis is frail and insufficient to find hidden paths in environment-intensive program. By analyzing a broad spectrum of different concolic testing systems, we conclude that a number of them cannot handle programs that interact with the environment or require a complete working model. This paper addresses this problem by automatically identifying and modifying outputs of the data input interface function(DIIF). The approach is based on fine-grained taint analysis for detecting and updating the data that interacts with the environment to generate a new set of inputs to execute hidden paths. Moreover, we developed a prototype and conducted extensive experiments using a set of complex and environmentally intensive programs. Finally, the result demonstrates that our approach could identify the DIIF precisely and discover hidden path obviously.