Previous studies fail to consider the importance of heterogeneity in the relationship between conflict and trade openness. In this paper, we fill this gap by using quantile regression to analyse the impact of conflict types (internal, external and global) on trade openness (de jure and de facto) in African countries over the period 1984–2017. The quantile regression approach thus enables us to identify the different responses of trade openness to conflicts at different levels of the conditional distribution of trade openness. Overall, our results show that conflicts reduce trade openness, and the impact is greater in countries with a low degree of openness. In particular, the effect of conflict types is negative, and its magnitude decreases as the level of de facto trade openness increases. On the other hand, this negative effect remains stable over the entire conditional distribution of de jure trade openness. Our results do not depend on the choice of conflict and trade openness indicators used. They suggest that countries with the lowest levels of openness are the most affected by conflict.