The study develops an intelligent decision-support system to quantify the inherent smartness of a traditional urban system. It strongly emphasizes the importance of integrating complex ecosystems of people, their institutions, and heritage and advocates transdisciplinary soft computational techniques to deal with present day challenges for resilient smart city planning. Four levels of abstraction are identified for the hierarchical evaluation of households’ responses within the delineated study area of Alwar walled city in north Indian state of Rajasthan. The residents’ perceptions are aggregated using fuzzy arithmetic averaging, which are further aggregated using fuzzy weighted averaging. Assuming this derived value as the response variable, a Sugeno fuzzy inference system is trained using leave-one-out-cross-validation (LOOCV) and optimized using metaheuristics algorithms. A genetic algorithm-based optimized model with urban fabric, network topology, and cultural vitality as independent input variables is proposed to predict inherent smartness quotient of an urban system. Application of partial compensatory approach, fuzzy linguistic variables for aggregation, and non-linear optimization enhance the readability, flexibility, and replicability of the model. The proposed model can provide an effective decision and planning support tool to policymakers and planners and facilitate benchmarking of spatial units to make informed choices, while considering the qualitative and quantitative parameters.