Background. This report describes the development and validation of a brain subscale for the Functional Assessment of Cancer Therapy (FACT) scale, and the revalidation of the subscales of the general version (FACT‐G), which measure physical, social, family, emotional, and functional well‐being and the quality of the relationship with the physician. Methods. 101 patients with primary brain tumors, after giving informed consent, participated in the last two phases of a four‐phase validation process: item generation, item reduction, validation, and reliability testing. In the validation phase, FACT‐G subscale and total scores as well as the brain subscale scores were correlated with other tests of mood, response, bias, and quality of life (QOL). Test‐retest reliability testing was performed with 46 patients who had primary brain tumors. Results. Validity and reliability coefficients were high for the FACT‐G and brain subscale, except for the comparison with a second QOL measure (FP‐QLI) and the Karnofsky Performance Status (KPS). The lower scores were the result of inherent differences in the two QOL instruments and the relatively high performance status of the brain tumor patients, which restricted the KPS score range. Conclusion. The FACT‐G has good psychometric properties supporting its broad generalizability and the brain subscale tests substantially different QOL issues than the core instrument. Use of this scale with the addition of the brain subscale provides a well rounded view of the various aspects of QOL from the patient's perspective. With modifications and further psychometric testing, the brain subscale may have broader applicability to subpopulations of patients with other brain disorders. Cancer 1995;75:1151–61.