The performance reliability of structural health monitoring methods depends on the accurate acquisition of a system’s dynamic responses and the consideration of uncertainties related to optimal sensor placement. This study presents a two-step framework for Near-optimal sensor placement and evaluates its robustness especially against uncertainties arising from potential damage scenarios, and modeling noise. First, the identification of the system using Covariance-driven Stochastic Subspace Identification method in the form of a finite element model and in offline conditions is performed and healthy and damaged modes are extracted under assumed scenarios. A novel index based on modal strain energy changes is then introduced to guide sensor placement. The effectiveness of the proposed sensor configuration is subsequently validated using a genetic algorithm and benchmarked against the optimal configuration. Then, the accurate structural damage identification process is employed by applying the Reference-based Stochastic Subspace Identification method to the incomplete data obtained from the proposed sensor placement configuration and combining it with the Probabilistic Strategy to estimate the probability of damage in the structural members. The proposed framework effectively balances accuracy, resilience to uncertainty, and computational efficiency, offering a practical and scalable solution for reliable sensor deployment and damage detection in real-world structural systems.