Summary Solution-based processing of photovoltaic materials has important advantages, including lower overall costs and easier upscaling of fabrication, where the solvents play an essential role. Antisolvent crystallization technique is often applied to achieve high-quality perovskite materials. However, the basic chemistry of the solvent-antisolvent crystallization is still not well understood. Besides, common antisolvents are toxic, and their number is limited. In this work, a novel robot-based method was used to screen the efficient antisolvents for different solvent-perovskite systems. 336 combinations of perovskite-solvent-antisolvent could be prepared and characterized by the integrated robotic platform in 2 days. In addition, we provided a detailed working mechanism of the solvent-antisolvent crystallization approach. Furthermore, hundreds of potential antisolvents were proposed based on high-throughput screening and simulation. Verification tests match well with theory, and all reported antisolvents used for photovoltaic device optimizations are within our predicted range of the Hansen space, indicating that the predictions are reliable.