As furrow irrigation is practiced on a large percentage of farm fields in Iran, the design and operation of this system calls for improvement. As collecting the necessary data to predict of advance and recession curves in a furrow requires field accurate measurements. These operations are costly and time consuming and the use of approved models is advantageous. The main objective of this study was evaluating the hydrodynamic, zero inertia, and kinematic wave models prediction by field observation using the SIRMOD software. Sensitivity analysis for this software was performed by employing three input parameters: discharge, Manning’s roughness coefficient and furrow slope. Field experiment performed at the Research Farm of Faculty of Water Sciences Engineering, the Shahid Chamran University of Ahvaz. Data were collected from three furrows, 60, 80 and 90 meters long, with three replications and three discharges of 1.0, 1.25 and 1.5 liters. sec-1. Four indices: 45 degree line (a), model average prediction error (Er), regression coefficient (R2) and average relative error of model (Ea) were used to evaluate the model’s predictions. The results of sensitivity analysis indicated that the SIRMOD software was sensitive to variations of input parameters. The predicted values of advance phase for all models were more than the observed values, but, generally, the amounts of (a) and R2 indicated that the hydrodynamic and zero inertia models had better performance in prediction of advance phase compared with the kinematic wave model. Moreover, the hydrodynamic and zero inertia models, as in the advance phase, predicted the regression phase better compared with the kinematic wave model.