Previous studies showed that housing conditions have a major impact on young horses’ behaviour: horses kept in group in paddock are less prone to express undesirable behaviours at weaning and at work than stalled ones. A study highlighted that the time spent stabled was associated with an increase of aggressiveness in horses. In riding schools, horse behaviour is crucial regarding users’ and professionals’ safety. Our study aimed to investigate how human management could impact on riding school horses’ behaviour. 3 emotionality tests (Arena, Novel object & Bridge tests), and 1 learning test (Chest test) were performed on 184 horses from 22 riding schools involved in the same general activity (teaching, beginner to moderate level) and differing in terms of housing conditions. Factorial correspondence and multivariate analyses were used to assess the effect of each factor on horses’ behaviour. Non parametric tests (Kruskall-Wallis & Mann-Whitney) were used to compare horses’ responses between categories of schools. The results show that riding schools could be separated into 4 categories according to their horses’ behaviour. The MANOVA revealed a strong impact of breed (Wilks’λ=0.49,F(60, 646)=2.14,P<0.001) and housing conditions (Wilks’λ=0.33,F(4,179)=7.56,P<0.001) on horses’ behaviour. In particular, horses from riding schools with box housing reacted more strongly in the novel object test (MW, Xbox=21.11±2.37, Xpadd=9.65±1.51,U=9,P<0.005) and showed more active locomotion (MW, Xbox=42.66±5.98% of orses/school, Xpadd=6.56±3.99% of horses/school, U=5.5,P=0.0005). This study highlights the importance for riding school owners to take into account horses individual characteristics, as well as to have a more general reflection on how management (in particular housing) may impact on safety.