The Innovation for Agribusiness (InovAgro) project, which launched with its first three year phase in 2010, uses a market system development (MSD) approach towards the goal of increasing incomes of men and women small-scale farmers in northern Mozambique. InovAgro interventions promote improved agricultural productivity, participation in selected high-potential value chains and the development of inclusive and sustainable market systems, such that impacts are expected to last long beyond the termination of the project. This paper presents results from a midline quantitative impact evaluation of the second phase of the InovAgro project interventions (2014-2017). In it, we use a carefully designed and executed quasi-experimental study design to credibly attribute changes in market engagement and welfare of participating farmers to exposure to the InovAgro II project, identifying and testing in what respects the intervention was most successful, and what regard it had less impact. Although InovAgro II projects operate in 11 districts of Zambézia and Cabo Delgado provinces, this impact evaluation focuses on two districts in Zambézia province (Alto Molócue and Molumbo), and in terms of value chains, focuses on the soybean and pigeon pea high-potential value chains, while the InovAgro II project interventions focus on these in addition to maize, sesame and groundnut. A baseline survey was undertaken in 2015 covering the 2014/2015 agricultural season and a midline follow-up survey was conducted in 2017, covering the 2016/2017 agricultural season and reaching 1,749 households of the original 1,886 households interviewed in the baseline survey. Using difference-in-difference estimation and propensity score matching, we find that exposure to the InovAgro II project is associated with an increase in the proportion of households selling soybean and pigeon pea by approximately 5% and 16%, respectively (significant at the .01 level). Exposure to the InovAgro II project also results in significantly higher shares of smallholder farmers using improved seed for soybean and pigeon pea (an increase of 6% for soybean and 2% for pigeon pea). We find that the InovAgro II project is also associated with significant increases in access to agricultural output market information from formal sources (5%) and hired labor for farming activities (8%). Despite the significant impacts on short term outcome variables, exposure to the InovAgro II project had limited impact on long term outcome variables, such as on rural-urban migration as well as engagement in the non-farm sector (two proxies for assessing potential welfare implications of the project) however this finding is not surprising given the impact evaluation covers only two years-a short period of time to bring about the long-term impacts expected to eventually emanate from an MSD project.