The economic literature on the spatial determinants of innovation has emphasized the fact that the environmentin which a firm is located influences its ability to innovate (Massard&Mehier, 2009; Boschma&Iammarino, 2009).This is highlighted through the concept of spatial externalities, whose properties depend on the characteristics ofthe areas in which the firms are located. Furthermore, the notion of agglomeration often leads to solely retain thebenefits of locating in a dense environment without taking into account the diversity of the territories and of thetypes of externalities they can generate (Torre&Wallet, 2014). Yet, territories also have rural areas characterizedby a lower density but also by different socio-economic structures and dynamics. Our objective is to analyze thedeterminants of the innovative performance of French industrial firms according to their urban/rural location,questioning in particular the role of spatial externalities in terms of specialization, diversification and relativevariety (Marshall, 1920; Jacobs, 1960; Frenken&al., 2007) and innovation models that are at the basis of smartdevelopment (Foray&al., 2011). The model built will articulate the firm's internal characteristics (size, multi-unit,...) and the spatial, sectoral and market dimensions of its environment. The first assumption is that both theinternal characteristics and the external environment play an important role in building the knowledge base of thefirm and its ability to innovate. The second assumption is that according to its location, the different types ofspatial externalities can play differentially on the innovative performance. Our empirical data are based onsurveys and zonings of the French Institute of Statistics. The Community Innovation Survey 2008 is used tocharacterize the firm’s processes of innovation. The Annual Survey of Firms 2007 provides general and locationalinformation about the industrial firms. The Declaration of Social Data 2008 is used for calculating the indices ofspecialization, diversification and related variety for employment areas (Galliano&al., 2014; Frenken&al, 2007). Forthe degree of urbanization, we mobilize the 2010 urban units approach based on density and the 2010 urbanareas zoning based on commuting rates and stock of employment. Then, we will use a selection model to analyseboth stages of the innovative process; the firms’ decision to engage in an innovative process and the intensity ofthe innovation activities. The results show the effects of firms’ individual characteristics on the benefits they getfrom their spatial location. They highlight that rural firms have forms of mobilization of external resources quitedifferent from those of urban firms, and that spatial externalities influence innovation in a differentiated manner.Rural firms are in particular more sensitive to Marshallian and related variety environments than urban firms moresensitive to diversified environments.