Building energy consumption, as an important part of the total social energy consumption that cannot be ignored, is having a far-reaching impact on energy consumption and environmental protection, and with the deepening of the urbanization process, the problem of energy management of the building complex, which is the core carrier of urban construction, is becoming more and more prominent. Based on the uncertainty inference model of Bayesian network, this paper proposes an intelligent energy management uncertainty modeling methodology system and constructs a Bayesian energy management network model with adaptive learning ability. Relying on the conditional probability inference mechanism and graph theoretic expression, the model can adjust the decision-making strategy in a timely manner in the complex and changing environment, which greatly improves the effect of energy consumption control of building complexes under uncertainty conditions. The introduction of Monte Carlo Markov chain and sequential Monte Carlo methods enhances the evaluation ability of various energy consumption management models through Bayesian data analysis, and ensures the reliability of the algorithmic model selection and comparison process. Aiming at the intricate correlation phenomenon among many variables in the energy consumption system of building clusters, a node-based hierarchical Bayesian network improvement structure is proposed, in which each building unit within a building cluster is regarded as a basic element of the energy management network, and a hybrid network construction method that combines static topology analysis and dynamic correlation optimization is adopted. Through the multi-level network division and node dynamic connection mechanism, the expression accuracy of the building cluster energy consumption correlation relationship is significantly improved. The study shows that the intelligent energy management method of building cluster based on Bayesian network effectively handles system uncertainty through probabilistic reasoning, realizes accurate assessment and prediction of energy consumption state, and can adjust decisionmaking strategies in real time according to environmental changes and energy consumption state. It improves the adaptability and efficiency of energy consumption management, which is expected to produce significant energysaving benefits and economic value in practical engineering applications.
