Abstarct: Side weirs are among the most important hydraulic structures which is used as water level controller and flow rate measurement in channels. Side weirs are widely used in hydraulic, environmental applications as well as irrigation and sewage treatment systems and water diversion from the main canals to the hydraulic lab and irrigation and drainage networks. Precision in measuring flow for different flow conditions has led to the design of various geometric shapes of side weirs. Therefore, due to the importance and extensive applications of lateral side weirs, these overflows must have properties in order to be able to properly utilize them in water transmission networks. One of the most prominent features that can be mentioned here is the accurate measurement of the flow through the side weir and its ability to drain significant water during flood event. Because simple overflows during flood events are suitable only for excess water flow in the canal or river, but not suitable for accurate measurement of passage. Therefore, the accuracy of measuring the discharge and water flow during flood events, is one of the most important advantage of composite side weirs. In this study, the rectangular-rectangular, triangular-rectangular, sharp, perpendicular, overheads are examined at 60, 90,120 degree angles. Therefore, predicting the intake flood discharge coefficient using regression analysis, improve analyse regression, artificial neural networks (GMDH) and Gene exxpression planning (GEP) are among the most powerful and must widely used modelers. Also, for the modeling, the effective parameters have been used for the flow coefficient. For comparison, statistical parameters such as R2 and RMSE have been used. The results indicate that GEP programming with the amount for training data and for test data in a rectangular-rectangular compound side weir and for triangular-rectangular composite side weirs of 60,90,120 degrees respectively for equal education data, and for the test data were equal, and obtained compared to the regression model, SPSS and GMDH, the RMSE error value for the rectangular-rectangular compound side weir test was 0.023, 0.0015, and 0.007 respectively, for a triangular-rectangular composite side weir with a 60 degree angle are 0.041, 0.002, and 0.005, respectively and for rectangular-triangular side weir an angle of 120 degrees, respectively equal to 0.004, 0.001, 0.009 percent, which indicates the superiority and optimality of the modeling of gene exxpression planning.