The study has adopted a knowledge-based, decision support system of Bayesian network (BN) modeling approach to address the lack of incorporation of the interlinkage of dimensions and uncertainty associated with neighborhood sustainability assessment (NSA). The study aims in developing a three-tier top-down BN model incorporating expert elicitation with three sub-models constituting 30 nodes concerning the Economic, Social, and Cultural Dimensions for the assessment of economic-sociocultural (ESC) sustainability of neighborhood-level urban communities (NLUCs) in Kolkata, an Indian megacity. The conceptual BN investigates the causal relationships and effects of different indicators and sub-models by quantifying their influence in achieving ESC sustainability. The study has proposed a robust algorithm for model parametrization and a detailed methodology for its structure development and validation. It has been tested using 550 sets of real-world survey data, with a prediction error rate of 2% for the query node. The model suggests the social sub-model is most sensitive to ESC sustainability, followed by the Economic and Cultural sub-models. It concludes with a discourse concerning spatial planning of NLUCs on dealing with the COVID-19 pandemic in compact cities of the Global South and discusses the relevance of the indicator set for achieving long-term sustainability.