This study examines reliability (test-retest and interrater) and predictive validity of the Home Assessment Profile (HAP), a quantitative, performance-based home assessment instrument. As part of a prospective cohort study of falls risk in older adults, the homes of 159 community-dwelling male veterans with impaired mobility were assessed at baseline using the HAP. Participants were then followed for 6 months for falls. The intraclass correlation coefficient (ICC 1,3) was used to assess agreement both between raters and between days for the total HAP score. The ability of the HAP to predict falls was assessed using logistic regression, adjusting for age, cognition, and mobility level. Total HAP scores ranged from 4 to 64 for the interrater trials and from 9 to 57 for the test-retest trials. Agreement between raters was 0.92 and between days was 0.92 for total HAP scores. Adjusting for age, cognition, and mobility level, the HAP was a significant predictor of falls. The HAP is a reliable and valid quantitative instrument that may be a useful adjunct to comprehensive geriatric assessment.