Abstract: The objective of this study was to investigate the effect of landscape complexity on biological control of the millet head miner Heliocheilus albipunctella de Joannis (Lepidoptera: Noctuidae) a key millet pest in West Africa. Field observations were carried out in 2013 and 2014, in a 20 × 20 km area around Dangalma village (14°43’42’’N, 16°33’98’’E), in the Diourbel region located in the main millet-producing basin in Senegal. Biocontrol was evaluated by excluding natural enemies to infested millet heads. Biocontrol Services Index (BSI) was calculated on a set of 45 millet fields separated by a distance of 2 km from each other. Landscape complexity around each field was assessed yearly using five landscape metrics, calculated from two land cover maps derived from Pleiades satellite images. To study the effect of landscape attributes on BSI values, a generalized linear model was performed at 9 different spatial scales (from 0.250 to 2.250 km around fields). The best statistical model was used to identify the environmental key variables enhancing biological control of the millet head miner (MHM). We found that BSI values increased with landscape diversity, measured as Shannon’s Diversity Index (SHDI). Moreover, landscapes dominated by millet fields provided less regulation of MHM than landscapes dominated by semi-natural vegetation. Landscape diversity and composition at a 1750 m around sampling sites explained the greatest proportion of the variation of biological control of the millet head miner. This study indicates that semi-natural vegetation, here mainly trees, have to be conserved to enhance biocontrol of H. albipunctella.