In the present study, attempts have been made to predict the soil infiltration rate by using the Kostiakov infiltration model. Subsequently, the feedforward backpropagation Artificial Neural Network (ANN) was employed to evaluate the constants of the Kostiakov infiltration model. The double ring infiltrometer approach was employed for collecting the infiltration test data. Infiltration tests were carried out for winter and summer seasons on 106 observation points, over the study area. The parameters of the infiltration model stated above are determined along with this data for different soil properties like bulk density; moisture content; % sand; % silt; % clay; electrical conductivity; field capacity; and wilting point which were determined by experimentation. These data serve as input to the ANN and the parameters of the Kostiakov model were determined. The performances of ANN models with the different input combinations are evaluated for the prediction of the Kostiakov model parameters on clay soil. The performance of the ANN was assessed based on various evaluation criteria. The Nash-Sutcliffe efficiency was observed to be 93.79% and 97.71% for model parameters ‘a’ and 'b' respectively. Thus, the study demonstrates the application of ANN for the evaluation of infiltration characteristics of clay soil obviating the need for carrying out tedious field experimentation.