In the context of implementing large-scale projects, such as the construction of nuclear power plants (NPPs), it is crucial to ensure the predictability, controllability, and optimality of construction plans. This article presents a generalized probabilistic project management model based on the combination of PERT, GERT, Monte Carlo methods, mathematical optimization, and modern machine learning and data analysis (Data Science) techniques (2018). The purpose of the article is to formalize an approach that allows for:• building predictive network models considering uncertainty,• identifying critical and subcritical paths,• performing local and global optimization of solutions,• ensuring model adaptation based on empirical data.