Conventional supply chains, involving several stages and various intermediaries, are affected by some well- known forms of inefficiencies and drawbacks. Besides the increase of market price consequent to multiple marginalization, such supply chains generally suffer of significant post-harvest losses and product waste. In this context, short food supply chains have been recently proposed as different systems capable of delivering higher quality products while promoting sustainability and efficiency. Across the EU, a growing number of consumers choose to buy food products on local farmers' markets, associating local products with higher quality standards (freshness, nutritional value), healthy eating, more environment-friendly production methods and lower carbon footprint. Such elements seem to confirm a higher performance of short food supply chains (SFSCs) compared to traditional (long) chains in terms of sustainability and quality of products. Nevertheless, the performance of SFSCs is significantly affected by the local contexts and the market situations in which they operate. In particular, although SFSCs are localized in relatively small geographical areas, the elimination of intermediaries, such as distributors/packagers, and quality preserving processes generally results in shorter shelf-life of products. The management of such systems, hence, is focused on the problem of ensuring superior quality of local product at reasonable costs, without the possibility of employing advanced packaging solutions. Therefore, due to their peculiarities, SFSCs require proper logistic policies to cope with these problems, taking into account the variability of demand and the effects of seasonality. This paper in particular focuses on the logistics of SFSCs and proposes a methodology for optimal inventory management, with the aim to preserve the shelf-life of the products, and to ensure supply chain efficiency. The methodology developed is based on a multi-objective approach to inventory management in a serial two-echelon system. A numerical application is proposed in order to prove the effectiveness of the model.