The alternative adopted in rice (Oryza sativa L.) breeding programs to decrease the effect of genotype x environment to reduce the effect of the interaction of genotype x environment (G x E), consists in the evaluation of genotypes in different environments and years, for an accurate recommendation of the most stable and adapted genotypes. Therefore, the objective of this work was to evaluate the phenotypic adaptability and stability of the productive upland rice genotypes in Mozambique, using the graphical tool for data analysis, the GGE biplot. We used data on grain yield, derived from the evaluation of five upland rice genotypes from the Rice Genetic Improvement Program of the Institute of Agrarian Research from Mozambique -Northeastern, Zonal Center (IIAM-CZnd). The trials were conducted in the complete block design with four repetitions in the agricultural years 2019/2020, 2020/2021, and 2021/2022. The trials were conducted in a total of five environments divided into three agricultural years, located in northern Mozambique. It was concluded from this work that genotypes G4 and G2 were superior to the overall average for productivity and genotype G3 showed high stability. The genotype G4 can be recommended for all mega-environments, for presenting good adaptability and phenotypic stability.