Human immunodeficiency virus (HIV) therapies need to be both effective and acceptable. The 10-item HIV Treatment Satisfaction Questionnaire (HIVTSQ) was validated amongst 150 HIV-1 sero-positive individuals, receiving one of two protease inhibitors as part of combined therapy in an open-label randomised trial. Scale and subscale scoring was determined psychometrically. It was hypothesised that satisfaction with control would be greater amongst those with lower viral loads, satisfaction with side-effects would be inversely related to severity of adverse events and satisfaction with the new treatment would be greater than with the control treatment. Principal components analyses suggested that patient ratings of nine items can be summed to compute the total satisfaction scale (Cronbach's alpha 0.82), and/or divided into subscales: general satisfaction/clinical (alpha 0.80) and lifestyle/ease (alpha 0.74). One item (asking how demanding the treatment was) needs modification before inclusion. The HIVTSQ showed construct validity: viral load correlated negatively (Spearman's r - 0.33 p < 0.01) with satisfaction with HIV control; those with <400 copies HIV-1 RNA/ml were more satisfied with HIV control than those with higher viral loads (Mann-Whitney p < 0.01); adverse event grade correlated r - 0.18 (p < 0.05) with satisfaction with side-effects. The HIVTSQ was sensitive to differences between groups: compared with patients in the control group, those receiving the new treatment had significantly higher perceived flexibility and lifestyle/ease scores at week 8 (Mann-Whitney p < 0.01). Patient perceptions did not simply mirror clinical measures, highlighting the importance of measuring patient views.