Ecosystem services (ES) concept is now used by many researchers. But some questions remain, especially on generalizability and mapping of results. Which method using? Some researchers suggest basing ES mapping on landscapes knowledge but, in fact, they only use land use map: a given ES is associated to a land use category. This is feasible for provisioning services but what about cultural services (CES)? This proposal intends to question this methodological point, in really different contexts, based on two researches. The one is about French urban green spaces (UGS), studied in the framework of a research project. The other one focused on sylvosystems of a Vietnamese rural area, in a PhD framework. In two cases, interviews were conducted, with users and managers of UGS, with Vietnamese villagers – completed by archives consultation. In France, the interviewees consider UGS as a source of well-being and consider them as leisure, relaxing or meeting areas. We distinguished different UGS: forest or semi-natural, ornamental and allotment gardens. We could think that ES could be associated to each one and thus mapping ES from UGS map. But the association is not so easy. Moreover, urban dwellers can consider a given category of urban nature sometimes as CES source, sometimes as disservices source. Generalizing the results is thus difficult and this compromises CES mapping. In Việt Nam, CES associated to sylvosystems also differ, here especially according to ethnic belonging of interviewees. Some consider forest as sacred, others with repulsion. The map could be based on localization of ethnic groups (to be considered for landscapes mapping). But, the sacred nature is not attached to all forests, depending on beliefs. Moreover, these CES varied also over time. The link between landscapes and ES is heuristic, because these concepts question the interrelations Man-Nature, in an interdisciplinary way. But, as it’s difficult to include the various assessments of landscapes, it’s difficult to generalize CES and thus to map them.