Innovation is widely recognized as being an important strategic tool for companies to increase their competitive advantage. Hereby, networks have become increasingly important as external sources for the necessary knowledge, ideas and financial resources. The main contribution of this paper is to shed light on how different network partners can explain or facilitate the different types of innovations in the agricultural sector. In contrast to other studies, we make a distinction between all four types of innovation: product, process, marketing and organizational innovation. Thus, this study has the objective to gain insight into the innovation process of farmers in terms of how they innovate, which network partners they consult in relation to innovation type, the obstacles they face, and where the network activities could be better aligned with the needs of the farmers, which could help to enable them to optimally support innovation and networking. The study is based on 36 in-depth interviews with farmers spread over five subsectors in Flanders (northern Belgium). Our most important findings are that the consulted partners and the observed barriers are different dependent on the innovation type. Hence, our study delivers a set of valuable insights and implications for farmers, network coordinators and policymakers. Farmers must be aware of the importance of partner suitability and network heterogeneity for the innovation type they are aiming at. Furthermore, farmers have to be aware of the fact that efficient networking is not the optimisation of single relationships independently of each other, but instead the management of synergies and coordination of all relationships in an efficient way. In addition, network coordinators should set up a clear strategy and communicate for which innovations their network can advise and help the farmer. These first conclusions should be further proven and supported by future research in order to draw general conclusions for the agricultural sector. As the sample of our study is limited to 36 respondents spread over five subsectors, it is necessary to conduct a quantitative study to achieve a representable sample and to include more subsectors. In addition, the study is limited to the Flemish region and literature in other countries about this subject is scarce. Hence, other researchers are encouraged to investigate if the results of Flanders can be supported by other regions in Europe and the world.