Discusses the application of Bayesian Networks in the analysis of risks in the Eddy-Current (EC) inspection of critical parts/pieces. As a methodological approach, Bayesian Belief Networks (BBN) was chosen as a tool to predict the most significant risks and the best measures to avoid them. The process was entirely mapped in order to identify the factors with higher probability of resulting in failure when executing the inspection. The results indicate that activities with higher chance of failure during inspection are related to labor factors and equipment. The conclusion is that the proposed method is crucial for correct identification of critical factors and most effective countermeasures, informing for decision makers how to avoid or mitigate errors in the inspection and, consequently, failures in critical components.