For patients diagnosed with chronic illness, attitude towards treatment may play an important role in health and survival. For example, negative attitudes towards treatment have been related to poorer adherence to treatment recommendations and prescribed medication across a range of chronic illnesses. In addition, prior research has shown that attitude towards treatment assessed through a psychiatric interview predicted survival at 1year after bone marrow transplantation with great accuracy (> 90%). The purpose of this study was to determine the relationship between a self-report attitude to a treatment measure that operationalized a psychiatric interview, and survival over 17years in a sample of people living with HIV (PLWH). Participants (N = 177) who were in the mid-range of HIV illness at baseline (CD4s 150 to 500, no prior AIDS-defining clinical symptom) were administered the Montreal-Miami Attitude to Treatment (MMAT-20/HIV) scale and followed longitudinally to determine survival at 17years. The Montreal-Miami Attitude to Treatment (MMAT-20/HIV) scale is a 20-item self-report questionnaire designed to survey multiple factors that contribute to an overall psychological construct of the treatment process. The MMAT-20/HIV predicted survival over 17years controlling for biomedical (baseline CD4, viral load, antiretroviral medications, age) and psychosocial (race, education, antiretroviral medications) variables. Those in the top half on the MMAT-20/HIV were almost twice as likely to survive than those in the lower half. Scores on the MMAT-20/HIV were significantly but modestly correlated with adherence (r = .20, p < .05), but adherence was not a mediator of the relationship between the MMAT-20/HIV and survival. An individual's attitude towards the treatment process predicted survival, raising the possibility that optimal clinical management would include ways to probe these attitudes and intervene where possible. The ease of administering the MMAT-20 and adaptability to other illnesses could facilitate this endeavor.