This paper presents a robust approach to develop ground thermal loads to be used for the long-term analysis of ground-coupled heat exchangers. The thermal loads developed using the proposed approach are simple, yet accounts for the hourly variation of the thermal loads. The analysis to develop these ground thermal loads is based on the change-point statistical analysis to determine the heating and cooling episodes. Heating and cooling loads are then converted into sine waves with an equivalent amplitude and the episodes determined from the change-point statistical analysis. Using the approximated equivalent sine waves, finite element or finite difference methods can be utilized to predict the long-term performance of the heat exchangers. The proposed methodology is compared to other existing approaches and showed better representation of the varying hourly ground thermal loads for a 1 year period. Moreover, a long-term thermal performance is predicted using the equivalent sine waves developed for balanced and unbalanced ground loads. (C) 2014 Elsevier Ltd. All rights reserved.