Assessing substance use of homeless persons is a critical task. This study examines the test-retest reliability, concurrent validity, and sensitivity to change of the Time-Line Follow-Back interview, a calendar instrument used to assess days and quantities of alcohol use and days of illicit drug use, in the multisite Collaborative Program to Prevent Homelessness (CPPH). The Time-Line Follow-Back was reliable for assessing use during the past month and the recent 6 months. Results from the Time-Line Follow-Back were correlated with other self-reports of use, with research diagnoses of substance use disorder, and with clinician ratings of severity of substance abuse. The Time-Line Follow-Back detected changes in clients with severe mental illness and in those with less severe psychiatric problems.