Numerical simulations have been widely employed in buildings, particularly for prediction of natural ventilation. The Computational Fluid Dynamics (CFD) has been utilized successfully for prediction of the airflow and heat transfer in solar chimney, one of the common devices for natural ventilation. However, CFD requires expensive computational resources. In this study, an Adaptive Neuro-Fuzzy Inference System (ANFIS) model was tested for quick prediction of the heat transfer in a vertical solar chimney based on the data provided by a CFD model. The Nusselt number in solar chimneys at different heights and heat flux were predicted with the CFD model. The generated data were used to train two ANFIS models which were then validated with the remaining CFD data. The results show that the ANFIS models can predict the Nusselt number and mass flow rate with the maximum discrepancy between the two results of less than 5.0%