Abstract Aim: To test the validity and reliability of a Swedish version of the Non-Communicating Children’s Pain Checklist – Postoperative Version (NCCPC-PV). Methods: Thirty-two consecutive children/adolescents (2–20 years of age) with cognitive impairment and no verbal communication from four habilitation centres were admitted to the study. Each child’s behaviour was observed by a parent or a caregiver and by a physiotherapist in two calm and two painful situations within the child’s everyday life. The raters independently assessed and graded the child’s behaviour during 5 min according to the translated Swedish version of the NCCPC-PV. The intrarater and interrater reliability were determined, and the construct validity was examined. Results: The results from 202 assessments showed that the construct validity was good: children’s behavioural signs differed significantly between situations of pain and situations of calm (p < 0.001). Repeated assessments showed poor agreement both within and between raters [intraclass correlation coefficient (ICC) 0.51–0.65]. The agreement for pain was good (ICC 0.83). Conclusion: The Swedish version of the NCCPC-PV can be used for pain assessment in children with cognitive impairments who lack verbal communication. Aspects of reliability need to be further analysed.