A provenance trial of Michelia chapensis Dandy in the south of China was carried out in three sites to assess yield and stability analyses. The trials were conducted as randomized, complete block designs with four replications at each site. Additive main effects and multiplicative interaction(AMMI) and genotype main effect and genotype 9 environment interaction(GGE) were employed in the evaluation of provenances; AMMI analyses showed that significant genotype 9 environment interaction effects(P 0.05) existed between provenances, capturing 15.11 % of the total sum of squares. The AMMI stability value revealed that provenance 2 was stable, but had low yield. GGE-biplot models showed that provenances 7, 8, 10, and 11 were more stable and highyielding. The GGE results also confirmed that among the three sites, site Jiu Qu Shui was the most ideal environment for representation and discrimination. A provenance trial of Michelia chapensis Dandy in the south of China was carried out in three sites to assess yield and stability analyzes. The trials were conducted as randomized, complete block designs with four replications at each site. Additive main effects and multiplicative interactions (AMMI ) and genotype main effect and genotype 9 environment interaction (GGE) were employed in the evaluation of provenances; AMMI analyzes showed that significant genotype 9 environment interaction effects (P 0.05) existed between provenances, capturing 15.11% of the total sum of squares The AMMI stability value revealed that provenance 2 was stable, but had low yield. GGE-biplot models showed that provenances 7, 8, 10, and 11 were more stable and highyielding. The GGE results also confirmed that among the three sites, site Jiu Qu Shui was the most ideal environment for representation and discrimination.