This paper explores the integration of Bayesian Networks (BNs) and Mixed-Effect Models to analyze local well-being indicators within the Equitable and Sustainable Well-being (BES) framework. By extending Conditional Gaussian Bayesian Networks (CGBNs) with random effects, our approach explicitly incorporates both spatial and temporal dimensions, allowing for improved estimation of missing data and capturing local heterogeneity. The results highlight the superiority of mixed-effects models with partial pooling over traditional BN approaches, demonstrating their effectiveness in incorporating spatial and temporal variability. The study provides insights into how well-being indicators evolve over time and vary across different regions, offering a comprehensive framework for policy analysis and decision-making. Future research will focus on improving predictive accuracy, exploring spatial autocorrelation, and extending the methodology to dynamic Bayesian networks.