BACKGROUND: HIV testing and serostatus awareness are essential to implement biomedical strategies (treatment as prevention; oral chemoprophylaxis), and for effective serostatus-based behaviours (HIV serosorting; strategic positioning). The analysis focuses on the associations between reported sexual risks, the perceived risk for HIV infection, and HIV testing behaviour in order to identify the most relevant barriers for HIV test uptake among MSM living in Germany. METHODS: MSM were recruited to a nationwide anonymous online-survey in 2013 on MSM social networking/dating sites. Questions covered testing behaviours, reasons for testing decisions, and HIV risk perception (5-point scale). Additional questions addressed arguments in favour of home/ home collection testing (HT). Using descriptive statistics and logistic regression we compared men reporting recent HIV testing (RT; previous 12 month) with men never tested (NT) in a subsample not previously diagnosed with HIV and reporting ≥2 episodes of condomless anal intercourse (CLAI) with a non-steady partner of unknown HIV serostatus in the previous 12 months. RESULTS: The subsample consisted of 775 RT (13 % of RT) and 396 NT (7 % of NT). The number of CLAI episodes in the last 12 months with non-steady partners of unknown HIV status did not differ significantly between the groups, but RT reported significantly higher numbers of partners (>5 AI partners: 65 vs. 44 %). While perceived risks regarding last AI were comparable between the groups, 49vs. 30 % NT were <30 years, lived more often in towns/villages <100,000 residents (60 vs. 39 %), were less out-particularly towards care providers-about being attracted to men (aOR 10.1; 6.9-14.8), more often identified as bisexual (aOR 3.5; 2.5-4.8), and reported lower testing intentions (aOR 0.08; 0.06-0.11). Perceived risks (67 %) and routine testing (49 %) were the most common testing reasons for RT, while the strong belief not to be infected (59 %) and various worries (41 %) and fears of testing positive (35 %) were predominant reasons of NT. Greater anonymity (aOR 3.2; 2.4-4.4), less embarrassment, (aOR 2.8; 1.9-4.1), and avoiding discussions on sexual behaviour (aOR 1.6; 1.1-2.2) were emphasized in favour of HT by NT. CONCLUSIONS: Perceived partner knowledge and reasons reflecting perceived gay- and HIV-related stigma predicted testing decisions rather than risk perception. Access barriers for testing should be further lowered, e.g. by making affordable HT available, addressing structural barriers (stigma), and emphasizing beneficial aspects of serostatus awareness.