This article presents a methodology for risk assessment in complex projects, especially in highly uncertain environments. It highlights how emerging risks and interdependencies can make management in these projects difficult, requiring advanced tools such as the Bow-Tie model and Bayesian networks. The Bow-Tie model allows a structured visualisation of the causes and effects of the main risks, while Bayesian networks offer a probabilistic structure that facilitates the detailed analysis of complex scenarios. A practical example programmed in Python is included, in which a Bayesian network is used to model these risks and obtain a more precise and dynamic assessment, adapted to the uncertainty inherent in complex systems.