Background Although many within-subjects comparisons conducted on samples of men who have sex with men have sought to understand the association between specific situational characteristics (e.g. drug use or location of sex) and sexual risk behaviour, none have considered the ‘clustering’ of patterns of situational characteristics. An initial typology of sexual encounters is derived and the relationship of this typology to condomless anal intercourse (CAI) and pleasure is tested. Methods: Data from a longitudinal survey of men who have sex with men living in England were used. Multilevel latent class analyses were estimated to determine an optimal class solution on the situational characteristics, and then pseudo-imputation was used to estimate the association between class and both CAI and pleasure. Results: A three-class solution fit the data best, with a scaled relative entropy of 92.4%. Classes were characterised as featuring: regular steady partners in private locations with low drug use (class 1), casual partners with increased probability of sex occurring in a sex-on-premises venue (class 2), and high levels of polydrug use together with increased probability of casual partners (class 3). Encounters were different both in pairwise comparisons and overall on probability of CAI. They were different overall but not necessarily pairwise on pleasure. Conclusions: These initial findings demonstrate the possibility of understanding sexual encounters in terms of the contexts, or classes, within which they occur. This may have implications for tailoring HIV prevention to specific encounter types. Future research should seek to extend encounter-level typologies to specific drug use variables.