Selecting services from those available according to user preferences plays an important role due to the exploding number of services. Current solutions for services selection focus on selecting services based either only on non functional features, on context preferences or profile preferences. This paper discusses an improvement of existing services selection approaches by considering both user context and profile. The ultimate aim is to derive maximum profit from available profile and context information of the user by inferring the most relevant preferences w.r.t his/her contextual profile. Linguistic/fuzzy preference modeling and fuzzy inference based approach are used to achieve efficiently a selection process. Some experiments are conducted to validate our approach.