Abstract: OBJECTIVE: Prescribing of ADHD medications has not been examined in the US homeless population. This study examined frequency of prescriptions for stimulant and non-stimulant ADHD medications, as well as risky and potentially inappropriate prescribing (RPIP) of stimulants, in three groups of veterans with mental illness. METHODS: Using 2021-2022 national VA administrative data, we compared frequency of stimulant and non-stimulant ADHD medication prescriptions using logistic regression between homeless veterans (n = 105,062), veterans in the Department of Housing and Urban Development-VA Supportive Housing (HUD-VASH; n = 33,884), and independently housed (IH) veterans (n = 1,875,083). We also compared indicators of RPIP of stimulants using chi-square tests between the three groups. RESULTS: Adjusted for sociodemographic, clinical, and health care utilization characteristics, homeless veterans were less likely to be prescribed stimulants (adjusted odds ratio (aOR) = 0.80, 99 % CI = 0.76-0.85) relative to IH veterans and more likely to be prescribed non-stimulants (aOR = 1.12, CI = 1.06-1.18). However, among veterans prescribed stimulants, homeless veterans had more indicators of RPIP, including being prescribed stimulants in the presence of a psychotic disorder (7.9 % vs. 6.4 % for HUD-VASH vs. 2.2 % for IH, p < .001). CONCLUSION: Homeless veterans with mental illness were less likely to be prescribed stimulants and more likely to be prescribed non-stimulant ADHD medications relative to IH veterans with mental illness. However, RPIP of stimulant prescriptions was more common for homeless and HUD-VASH veterans relative to IH veterans.