The purposes of the study were to measure adherence with antiretroviral therapy to dose, schedule, and dietary instructions in a sample of patients with HIV infection in Stockholm, Sweden, over a 2-year period and identify baseline predictors of the three types of adherence. The study cohort consists of 144 patients who completed at least six out of seven follow-up self-reported adherence questionnaires. Baseline self-administrated questionnaire examined socio-demographics, medication-related, psychological, cognitive, and social context factors and self-reported adherence. Biomedical data were obtained through patients' medical records. Summary dose, schedule, and dietary instructions adherence scores provided outcome measures reflecting 100% adherence across all time points or not 100% adherence during at least one measurement period. A total of 61% maintained consistent full-dose adherence throughout baseline and all follow-up visits and equivalent proportion of 100% schedule adherence was 39%. Among patients with dietary instructions, 37% retained consistent adherence at all visits. Only schedule adherence was predicted by baseline data; perceived pressures from those close to the patient to take HIV medications (OR 0.51, p<.05), life stress (OR 0.13, p 0.009), ART health concerns (OR 0.19, p 0.003), and ART prolongs one's life (OR 0.39, p 0.04) predicted reduced schedule adherence over time. Perceived medication pressures from medical staff (OR 1.76, p<.05), post-traumatic stress disorder symptoms (OR 1.07 p<.01), and adherence self-efficacy (OR 3.50, p<.05) predicted positive schedule adherence over time. These results clearly illustrate difficulties in sustaining ART adherent behaviour, in particular schedule and dietary restrictions, over time and thus emphasizes the importance of multiple periodic assessments of all three types of adherence. Interventions aimed at improving schedule adherence should in particular focus on psychological and cognitive factors.