We conducted a discrete choice experiment (DCE) and quantified preferences for HIV testing among South African youth (Nov 2018 to Mar 2019). Six attributes and levels were identified through qualitative methods: source of HIV information; incentive amount and type; social support; testing method; and location. Each participant chose one of two options that comprised six attributes across 18 questions. Conditional logistic regression estimated the degree of preference [β]. Of 130 participants, median age was 21 years (interquartile range 19-23 years), majority female (58%), and 85% previously tested for HIV. Testing alone over accompanied by a friend (β = 0.22 vs. - 0.35; p < 0.01); SMS text over paper brochures (β = 0.13 vs. - 0.10; p < 0.01); higher incentive values (R50) over no incentive (β = 0.09 vs. - 0.07; p = 0.01); and food vouchers over cash (β = 0.06 vs. β =  - 0.08; p = 0.01) were preferred. Testing at a clinic or home and family encouragement were important. Tailoring HTS to youth preferences may increase HIV testing.