Considering the significance of the infiltration process in the design and management of surface irrigation and the artificial neural networks method capability in computational analysis of some processes, this research aimed to study the feasibility of using the artificial neural network method for detecting the main components affecting the infiltration of irrigation furrows. Infiltration experiments have been performed using a blocked furrow method. In applying artificial neural networks, the set of inputs including opportunity time, initial soil water content, flow depth, flow section area, wetted perimeter and wet bulk density, were considered and after mapping the data and selecting the appropriate hidden layer and using multilayer perceptron algorithms and principal component analysis for the training process, the cumulative infiltration values were satisfactorily estimated. The results indicated that the artificial neural network with the principal component analysis with a hidden layer and r = 0.97 and MSE = 0.006 in the validation phase is an appropriate method to analyze the infiltration of the furrows. Also, opportunity time and flow section area components, effectively influenced the cumulative infiltration of irrigation furrows.