Amendment of biochar (BC) in soil is an efficient and popular way to enhance agricultural productivity. In recent times, BC obtained from the by-products or wastes is gaining recognition in various engineering applications. Water hyacinth (WH), which is a highly invasive weed, can be turned into biochar for its productive usage. Recent studies investigated WH BC composite’s cracking potential, water retention capacity and agricultural productivity but did not focus much on its infiltration characteristics. The objective of this study is to investigate dependence of infiltration rate on crack intensity factor (CIF), suction and volumetric water content (VWC). The experiments were performed on the samples of bare soil, 5% and 10% BC (by weight) composites for 63 days (9 drying-wetting cycles) in natural conditions. The experimental data was used to train artificial neural networks (ANN). An ANN model was developed to predict the infiltration rate for each soil composition. Infiltration rate was relatively lower in case of 10% WH BC soil composite. CIF played a major role in governing the infiltration rate for bare soil but its significance relatively reduced as the BC content increased. BC content increases the relative importance of VWC in prediction of infiltration rate. Suction’s role in predicting infiltration rate, for both bare soil and BC composites was more or less the same. For applications (such as slopes or landfill cover) desiring less infiltration rate with a constraint of practically non-varying moisture content, 10% WH BC composite was found to be an ideal choice.