Recent research has analyzed whether higher levels of farm production diversity contribute to improved dietary quality in smallholder households. We add to this literature by using different indicators, thus testing the robustness of previous findings and helping to better understand the underlying linkages. The analysis builds on data from Indonesia, Kenya, and Uganda. Farm diversity measured through a simple species count has a small positive effect on dietary quality, either expressed in terms of dietary diversity scores or micronutrient consumption levels. However, when measuring production diversity in terms of the number of food groups produced, the effect turns insignificant in most cases. Further analysis suggests that diverse subsistence production contributes less to dietary quality than cash income generated through market sales. Much of the food diversity consumed in farm households is purchased from the market. If farm diversification responds to market incentives and builds on comparative advantage, it can contribute to improved income and nutrition. This may also involve cash crop production. On the other hand, increasing the number of food groups produced on the farm independent of market incentives will foster subsistence, reduce cash incomes, and thus rather worsen dietary quality. We conclude that from a nutrition perspective improving market access is more important than farm diversification as such.