Load prediction is the first step in designing an efficient community district heating system (CDHS). Even though several methods have been developed to predict the heating demand profile of buildings, there is a lack of method that can predict this profile for a large-scale community with a numerous user types in a timely manner and with an appropriate level of precision. This paper, first briefly describes the 4-step procedure developed earlier, utilizing a Multiple Non Linear Regression (MNLR) method, for predicting the heating demand profile of district, followed by description of the community structure, and its district system. It also reports the field measurement procedure for collecting the data required and the preliminary analysis data. Results obtained from a continuous monitoring of the CDHS over a two-year period is employed to validate the accuracy of the developed model in the predicting the CDHS's heating load profile. Finally, using the 4-step procedure, the district's energy demand profile is predicted, and compared with both the measured data and the initial prediction. The outcome shows a less than 11.2% in the mean square root error (MSRE) of the predicted and measured load profiles. (C) 2017 Elsevier Ltd. All rights reserved.