A better knowledge of the foraging ecology of predator insect species that feed on crop pests is of primary importance to improve their beneficial influence for agriculture and for the environment. Multi-agent models are an efficient tool in this respect since they can cope with the complex processes involved in individual behaviour in a heterogeneous space. We applied this method to model the behaviour of Episyrphus balteatus (De Geer, 1776), a helpful species of Syrphidae (Insecta, Diptera) which can overwinter as fertilized adult females and whose larvae feed on aphids occurring on both natural vegetation and crops. The “Hover-Winter” model focuses on the winter dynamics of an E. balteatus population at the landscape scale. Each individual is modelled as an autonomous agent who behaves according to a set of rules for foraging in the landscape, feeding on flowers, sheltering in forest edges and dying, constrained by climate and land cover. The model was developed from data in the literature, expert knowledge and field measurements. This paper presents the structure of the model, the definition of its parameters, the sensitivity of the model to their accuracy and its main outputs. Hover-winter is the first individual based model for E. balteatus and its preliminary results show that it helps to understand how E. balteatus uses agricultural landscapes to survive the winter.