Production of major global commodities like cacao, coffee, and palm oil largely depends on smallholder farmers that have little access to inputs. Yields lag far behind from what is theoretically possible, and climate change is expected to strongly influence future production. The perennial nature of these crops implies that initial decisions on cropping system design will determine crop yields over decades. Crop growth models for perennials are much-needed to design sustainable, climate-smart production systems, because long-term empirical data on the physiology, growth and yield of perennials are scarce. Compared to annual crops, model building for perennials faces specific challenges. Perennials develop over multiple years, thus effects of climate, and management practices like fertilizer application, can have long-lasting effects on growth and yield (van Vliet and Giller 2017; Woittiez et al. 2017). Effects of specific management practices, like pruning, have also to be taken into account. We are developing and updating process-based models for cacao and oil palm to improve yield projections. We will discuss ongoing model development and general challenges associated with building crop models for perennials, and we will provide an outlook to future model development.