In order to modeling and predicting heat transfer coefficient in na nofluids, art ificial neural network (ANN) and Adaptive Neuro-fuzzy Inference system (ANFIS) were used in this study. In ANN and ANFIS, Input data are Reynolds number and nanoparticles volume fractions, and output data is heat transfer coefficient. Both of them could predict very well, and there is good agreement between experimental data and predicted data. In ANFIS coefficient of determination (R2), average relative error and mean square error for train data are 0.99, 8.9×10–5 and 6.5476×10–5, respectively, and for test data are one, zero and zero. According to results, by increasing the Reynolds number and volume fractions, the heat transfer coefficient increases. For base fluid in Re = 16300, heat transfer coefficient is 10961.38 W/m2K, and for volume fraction 0.135, heat transfer coefficient is 13947.72 W/m2K, therefore, heat transfer coefficient of nanofluids increased 1.27 time compared to that of base fluid. Results obtained from ANFIS are reliable, and can be used in prediction. Also, for ANN, ARE, MSE and R2 value are, –0.003, 6.38264×10–5 and 0.99, respectively. So, there is good agreement between experimental data and ANN results too. According to errors, can conclude ANFIS is slightly better than ANN.