The forward-looking loan loss provisions management can smooth the pro-cyclical impact of bank credit supply on macro-economy, but there are few empirically estimable loan loss provisions management approaches with dynamic and forward-looking characteristics, thus limiting the use and effect of this macro-prudential policy tool. In the case that the borrower's asset growth follows a mean reverting stochastic process, this paper derives the dynamic random motion of borrower's asset values and characterizes the credit risk factors and forward-looking provision requirement within the framework of structural credit risk model. Different from the structural model, the asset growth in this paper has the characteristics of periodicity and stationary, which is not only fit for the inherent requirement of the forward-looking provision management, but also contributes to the empirical estimation and implementation promotion of the model. Based on the assets and liabilities data of industrial sector from 1993 to 2013, this paper also empirically estimates future credit risks and forward-looking provision requirement of the industrial sector and its sub-industries. Meanwhile, it also estimates the forward-looking provision component of regulatory requirement and that of the actual provisions in the banking system based on the following two official documents:PBOC[2002] No.98 and CBRC[2010] No.98, and judges their appropriateness combining with the portfolio status of borrower target leverage and financing costs based on the theoretical model. This paper deeply reveals the credit risks and their evolution trends of the industrial sector at all levels under different leverage and financing costs, and answers the hot issues such as the appropriateness of forward-looking provision requirement of the actual situation in the banking system and that stipulated by the regulatory authorities.