Modal property of structural system gets affected not only due to the presence of damage but also due to a variation of environmental agents like temperature, humidity etc. Detection of damage through modal parameter comparison thus may lead to false predictions. This article presents a two-stage data-driven approach in which damage detection and localization are performed in consequence. For detection, an auto-associative neural network (AANN) has been developed and its prediction error is defined as the novelty index. Each unique damage pattern has been observed to demonstrate a unique pattern in the prediction error and this pattern is also found to be consistent under different ambient temperatures. This prediction error is, therefore, modelled using a second artificial neural network (ANN) with radial basis function (RBF) as its activation function. Through this second ANN, the damage cases are classified against associated patterns in AANN prediction error and subsequently used for damage detection purpose. A numerical experiment is performed on a 2D truss bridge for validating this proposed algorithm. The results demonstrated the efficacy of this algorithm to be employed for damage detection under varying temperature.