Understanding surface currents is one of the fundamental topics in oceanography. These currents are caused by various factors such as the presence of prevailing wind, earth  rotation and changes in the density of seawater. In this study, the surface currents of the Strait of Hormuz have been predicted using the Group method of data handling (GMDH) model. To set up this model, the time series data of the surface flow velocity of this basin with an average time step of 5 days have been used. To investigate the surface areas affecting the currents of the Strait of Hormuz, the Persian Gulf basin was divided into 11 areas. Then, using regression analysis , GMDH model and 11 identified area, three marginal areas of the Strait of Hormuz that most affect the surface currents of this strait were selected. Then, using the water exchange model of Jones et al., Presented in 2003, different compounds were extracted from the three identified areas and used as input to the GMDH model. The results showed that the combination of these three zones and using of Jones flow pattern reduce the prediction error index compared to their pre-combined state. Then, due to the presence of anti-clockwise eddy on a basin scale in the months of March to July, the efficiency of the model was investigated by removing these months, which reduced the accuracy of the model with this method. Finally, by data classification, it was found that the GMDH model has good accuracy in predicting high speeds.