Solar collector, as the key component of any solar system, has always been the focal point of research in the field of solar energy. Based on the literature, data-based methods, which have been proven to be promising in accurate modelling of solar collectors, have not been used for modelling heat pipe solar collectors (HPSCs). At the same time, accurate equations relating the thermal efficiency of solar collectors to the operational and climatic conditions have not been obtained for HPSCs. Therefore, in this study, various data-based and energy balancebased modelling methods were proposed, and based on different accuracy criteria, their precisions were compared in predicting the performance of HPSCs. First, an experimental rig was manufactured and the operational data of the system was recorded throughout a year. The recorded experimental data was used to train and validate various modelling approaches. Then, the accuracies of the proposed models were analysed and assessed. The evaluated models included Artificial Neural Network (ANN), Thermal Resistance Network (TRN), Artificial Neuro Fuzzy Inference System (ANFIS), and Fuzzy methods. Among different modelling approaches, ANN had the best performance which was followed by the ANFIS and TRN methods. The Fuzzy method was not recommended due to its poor accuracy. In addition, the optimum equations relating the outlet temperature of HPSCs to the operational conditions of the solar water heating systems as well as the climatic conditions were obtained and verified.