The ability of cereal-legume intercropping to increase and stabilise yields in tropical climates has not been tested over a wide range of environments. We evaluated the STICS crop model and simulated a 20-year historical climate series for sole crops and cereal-legume intercropping on sites with contrasting climates in tropical regions. During the evaluation, the model simulated yields accurately except in the case of very low-yielding legumes in intercropping. The simulations then suggested that intercropping increased protein productivity and yield stability compared with cereal sole cropping, although these benefits diminished as nitrogen fertilisation increased. Recently published work on improvements to the model will enable these promising results to be consolidated.