Fast and accurate calculation of the predicted results of Chemical Vapor Depositions (CVD) is very helpful to the high-throughput optimization of the CVD processes. In addition, robustness of the calculation process is important for automation of the optimization process. Therefore, we have developed a novel calculation method, by which robust and accurate calculations along with reduced computing cost were achieved, to reproduce deposition profiles in a macroscopic cavity (macrocavity). Boundary value problems for estimating diffusion-reaction equations by iterations of numerical integrations were changed into problems of finding the linear combinations consisted of a few functions. The coefficients of the linear combinations were optimized by Genetic Algorithms (GA). We could demonstrate the validity of the proposed method using various examples of the reaction mechanisms and conditions.