Numerous seemingly promising cerebroprotectants previously validated in rodents almost all have failed in stroke clinical trials. The failure of clinical translation strikes an essential need to employ more ideal animal models in stroke research. Compared with the most commonly used rodent models of stroke, non-human primates (NHPs) are far more comparable to humans regarding brain anatomy, functionality and pathological features. The aim of this perspective was to summarise the advantages of NHPs stroke models over rodents, discuss the current limitations of NHPs models, and cast an outlook on the future development of NHPs in stroke preclinical research.