Yam (Dioscorea spp.) is an important source of food for millions of people throughout the tropics. Yam yields are constrained by several factors such as low soil fertility and a poor response of the crop to nutrient supply. Here we present an experimental and modeling study performed to test the hypothesis that N leaching in the humid tropics is a key factor that explains the poor response of water yam (Dioscorea alata L.) to N fertilization. A field experiment was carried out to assess the impact of three levels of N supply (50, 125 and 200 kg N ha(-1)) on crop growth and soil N availability. The N fertilizer was split into two equal doses (i.e., 50%/50% split ratio), one applied just before emergence and the other a month after tuber initiation, which corresponds to the current management used by farmers. The CropSyst-Yam model was adapted to include a new soil-crop N component describing the effects of limiting N conditions on yam growth, and then to test different N fertilizer managements. The experimental results indicated that all variables linked to the soil-crop system were affected by the increase in N supply (soil N availability, crop biomass, leaf area and N uptake), and the model well described these impacts. Nitrogen leaching throughout the field experiment represented about one third of the total N supplied by each treatment. Observed and predicted data revealed that N leaching reduced yam growth under the N50 and N125 treatments by reducing soil N availability near the tuber initiation stage. This effect was not observed under the N200 treatment. Model results showed that N leaching and its negative impact near tuber initiation in N50 and N125 could be diminished by advancing the second N application by two or three weeks, and by using a split ratio of 70%/30%. Model results also indicated that this fertilizer strategy would even be suitable during very rainy growing seasons.