Rivers are exporting increasing amounts of nitrogen (N) to lakes, which is leading to eutrophication. However, the seasonality apparent in nutrient loading, especially in tropical areas, is thus far only partially understood. This study aims to better understand the seasonality and the sources of dissolved inorganic nitrogen (DIN) inputs from sub-basins to tropical lakes. We integrated existing approaches into a seasonal model that accounts for seasonality in human activities, meteorology and hydrology, and we applied the model to the sub-basins of a representative tropical lake: Lake Tana, Ethiopia. The model quantifies the river export of DIN by season, source and sub-basin and also accounts for open defecation to land as a diffuse source of N in rivers. Seasonality parameters were calibrated, and model outputs were validated against measured nitrogen loads in the main river outlets. The calibrated model showed good agreement with the measured nitrogen loads at the outflow of the main rivers. The model distinguishes four seasons: rainy (July–September), post-rainy (October–December), dry (January–March) and pre-rainy (April–June). The river export of DIN to Lake Tana was about 9 kton in 2017 and showed spatial and temporal variability: It was highest in the rainy and lowest in the dry seasons. Diffuse sources from agriculture were important contributors of DIN to rivers in 2017, and animal manure was the dominant source in all seasons. Our seasonal sub-basins and rivers model provides opportunities to identify the main nutrient sources to the lake and to formulate effective water quality management options. An example is nutrient application level that correspond to the crop needs in the sub-basins. Furthermore, our model can be used to analyse future trends and serves as an example for other large tropical lakes experiencing eutrophication.