The accelerated increase in motorcycles in the country has led to a gap in the knowledge and maintenance of these vehicles. This gap has generated a growing demand for repair and maintenance services and, in turn, has led to a lack of information and education among motorcycle owners about basic mechanics and essential maintenance procedures. Motorcyclists face difficulties in identifying and troubleshooting common problems due to an insufficient understanding of mechanical systems and a lack of accurate and accessible information on proper motorcycle maintenance, resulting in undetected mechanical failures, which may present themselves early as noises or alerts before becoming more serious problems. These undetected failures not only affect the performance and durability of the vehicle but can also put the rider’s safety at risk. Issues such as brake wear, transmission system degradation, or suspension problems may not be visible until they become critical faults. Thus, the development of an expert system is necessary for new users to know the problems through basic alerts and based on a series of questions to know where the main failure occurs and the percentage of the same, avoiding more complex situations and improving safety and efficiency in the maintenance of the same. In this context, Python is used as a programming language, the UPAFuzzySystems library, and the expert library. Forward and backward chaining is used to develop diagnostic questions. Bayesian reasoning is used to obtain probabilities based on user responses, generating effective answers and solutions for users based on statistics. On the other hand, fuzzy logic is used to improve the accuracy and efficiency of the expert system by addressing the uncertainty and imprecision of user responses, allowing for a more detailed and accurate assessment of potential motorcycle mechanical problems. Thus, we present a fuzzy Bayesian expert system to contribute to the predictive maintenance of a motorcycle.