No instrument exists that measures the individual needs of forensic mental health service users (FMHSUs). The aim of this study was therefore to develop a valid and reliable individual needs assessment instrument for FMHSUs that incorporated staff and service user views and measured met and unmet needs. The Camberwell Assessment of Need was used as a template to develop CANFOR. Consensual and content validity were investigated with 50 forensic mental health professionals and 60 FMHSUs. Both were found to be satisfactory. Concurrent validity was tested using the Global Assessment of Functioning and a five-point needs scale, and again was found to be satisfactory. Reliability studies were carried out with 77 service users and 65 staff in high and medium security psychiatric services in the UK. Inter-rater reliability, rating whether a need was present or not, was high for service users (0.991) and staff (0.998). Similarly high reliability was found for unmet needs (0.985 and 0.972, respectively). Test–retest reliability was found to be moderately high for service users (0.795) and staff (0.852) when ratings were made two weeks apart. Similar levels were found for ratings of unmet needs (0.813 and 0.699, respectively). The average interview time was 23 minutes. CANFOR has good validity and reliability, and is suitable for further testing with other service user groups.