A side effect of urbanization, urban heat island (UHI), is well known in increases of ambient air temperature. This increase further leads to a rise in indoor environment temperature, reduction of thermal comfort, increase of cooling demand, and heat related morbidity and mortality especially among vulnerable people such as the elderlies and those living in poorly ventilated buildings. Thus, it is imperative for cities to be empowered with predictive tools during extreme heat waves in order to be able to provide emergency plans. For this purpose, it is utmost importance to develop specialized tools to predict the indoor conditions based on the outdoor conditions recorded at the weather stations. In order to develop a reliable warning system artificial neural network (ANN) and regression method were proposed and tested for an indoor air temperature forecasting application with respect to neighborhood parameters. To find the most practical approach, a cross comparison of the models was conducted by two different levels of simulation in order to present the capturing and prediction performance of the developed models. In general, the ANN model showed better accuracy in predicting the indoor dry-bulb temperature while it was more complicated in implementation. (c) 2014 Elsevier B.V. All rights reserved.