Purpose - A growing body of research has addressed the relationship between CSR and innovation. However, research on this relationship is still at an embryonic stage, and there is a need to better understand and further analyze the complex relationship within the specific context of SMEs, especially since innovation is seen as key to their survival and success. This paper proposes to address this gap in small business literature by studying the relationship between CSR practices and exploration and exploitation. It raises the following question: are SMEs that explore and exploit differently distinguished by the extent of their adoption of CSR practices? Literature - Indeed, we chose to rely not on innovation outputs themselves, but on exploration and exploitation activities (March 1991). The literature has suggested that exploitation may be positively related to incremental innovation performance, and exploration positively related to radical innovation performance (Atuahene-Gima 2005). In fact, radical and incremental innovations can be viewed as outputs of exploration and exploitation respectively (Bierly and Daly 2007). And yet, some authors have suggested that firms that adopt a CSR strategy may introduce innovations that are more incremental than radical (López-Pérez, Perez-Lopez, and Rodriguez-Ariza 2007). In this paper we consider CSR through environmental and human sustainability components; five dimensions were so retained: (1) workplace, (2) environmental policies, (3) market policies, (4) communities policies and (5) company values (see European Commission, 2005). Thus, one may suggest that firms with different CSR practices (in terms of environmental and social practices) explore and exploit differently: SMEs that intensively adopt CSR practices may focus more on exploitation than on exploration. Methodology - A questionnaire was administered to CEOs of firms in the French region Rhône-Alpes. Our final sample includes 488 SMEs located in this region. Statistical analyzes based on a methodological chain of data analysis methods are used to identify homogeneous groups of SMEs with respect to their activities of exploitation and exploration, then to analyze the potential effects of a set of CSR practices on these innovation activities. So, we first classified the firms according to their innovation activities which allowed us to distinguish and identify four groups of firms: "exploitation", "exploration", "exploitation and exploration" and "neither exploitation, nor exploration". We then conducted a detailed statistical description of the contents of each group from "standard" characteristics and CSR practices. Finally, a barycentric discriminant analysis was performed to identify CSR practices that differentiate and well separate these groups. Results and implications - Our findings confirm that not all CSR practices lead to innovation activities, justifying the need to take into account the nature of involved practices. We indeed find that the low adoption of a range of CSR practices is typical of firms in the "exploration" group, while the "exploitation" group is characterized by a higher effort to develop CSR practices, however few in number. Furthermore, the results show that the high adoption of a large range of CSR practices is related to firms in the "exploration and exploitation" group. In fact, while the group of "exploitation and exploration" is more similar to profile "exploration" when considering only innovation activities, it is more similar to profile "exploitation" when considering the CSR practices adopted by SMEs.