Seed innovation is one major factor for improving agricultural productivity. For some self‐pollinated varieties, such as wheat, farmers have the option to buy certified seed from seed dealers or to use their own farm‐saved seed. Historically, farmers could use farm‐saved seeds for free, which led to reduced incentives to innovate for private breeding companies. In recent decades, several countries have established different royalty systems for farm‐saved seeds to favor research investment. We developed a theoretical model to compare these different systems. We compared six stylized systems by analyzing their impact on incentives to innovate, as well as production efficiencies at both the seed and agricultural production levels. Our findings indicate that royalty systems allowing for a certain proportion of farm‐saved seeds result in improved welfare. The systems that lead to the highest total welfare levels are those in which the royalty level on farm‐saved seeds is regulated. This includes systems where the royalty is either directly defined by a regulator (as in the French or UK systems) or imposed to match the royalty level of the certified seeds (as in the Australian system). The Australian system performs better under high research costs. Conversely, under low research costs, the best system is either the French or the UK system, depending on the relative cost of producing farm‐saved seeds versus certified seeds. In conclusion, it is possible to design efficient royalty systems to create and produce innovation, in a context where farmers can self‐produce this innovation.