Aim The Swedish National Patient Register offers unique opportunities for epidemiological research of paediatric ischaemic stroke. We aimed to validate the diagnosis of paediatric ischaemic stroke in the National Patient Register to ensure the quality of future research. Methods The PedStroke cohort consists of 1606 individuals aged <18 years with a diagnosis of paediatric ischaemic stroke (ICD‐10: I63, I64; ICD‐8 and 9: 433, 434, 436) in Swedish national health registers between 1969 and 2016. We selected 292 cases for validation by reviewing medical charts. Results In all, 277 of the 292 medical charts were received, of which 273 had enough information to qualify for review. The diagnosis was correct in 242/273 cases, yielding a positive predictive value (PPV) of 89% (95% confidence interval (CI): 0.85–0.92) for paediatric ischaemic stroke in the National Patient Register. After validation, seven cases of 222 with childhood stroke were re‐categorised to perinatal stroke, resulting in a total of 56 perinatal stroke cases. In the Medical Birth Register, 38 stroke cases were identified of which 37 had correct diagnosis, generating a PPV of 97% (95% CI: 0.92–1.0). Incorrect diagnoses decreased over time and the number of diagnoses confirmed by radiology increased correspondingly. Conclusion The National Patient Register is reliable for epidemiological research of paediatric ischaemic stroke because of its high PPV for this diagnosis.