This paper presents a method for determining and controlling variables which are critical to the desired duration of a project. The proposed method consists of four stages: 1) a Monte Carlo simulation of project scheduling using a dependency struture matrix (DSM), 2) the reduction of the number of input variables using an index based on the standardized regression coefficient, 3) the determination and regionalization of critical variables using a modified Monte Carlo Filtering (mMCF) method, and 4) an evaluation of risk of the project duration under the regionalized conditions. The proposed method was applied to three case studies. Results show that the method could be helpful in scheduling project to obtain the desired duration under uncertainty, identifying critical variables, and regionalizing the critical input variables.