Background and justification: Ethiopia is the third largest sorghum producer in Africa next to Nigeria and Sudan. Shortage of widely adapted and stable high yielding variety is one of the major bottlenecks for production and productivity of sorghum in the country. Grain yield performance is not the only parameter for selection as a genotype with the highest grain yield would not necessarily mean stable and adaptable across location and years. Eberhart and Russell (1996) and AMMI model could be the preferable tools to identify stable, high yielding and adaptable genotype (s) for wider or specific environments. Objectives: To identify stable high yielding sorghum varieties that could be adapted for wider and/or specific environments and make recommendations for further demonstration and production in the test environments and similar agro ecologies. Material and methods: A total of 21 released sorghum varieties and a local check were evaluated at three locations in western Ethiopia (Bako, Gute, Biloboshe) and two locations in eastern Ethiopia (Mechara and Mieso) in 2017 main cropping seasons. The trial was arranged in a randomized complete block design (RCBD) in three replications. Summary of results and application of the study: The combined analysis of variance revealed highly significant effect of environment and genotype by environment interactions for grain yield. This indicated that the tested varieties showed inconsistent grain yield performance across locations. Birmash variety gave the highest grain yield with average yield of 3.5 ton ha-1 but specifically adapted to Gute, Biloboshe and Mechara. Baji was the second high yielding variety with mean grain yield of 3.3 ton ha-1 and relatively with wider adaptability. The first two IPCAs accounted for a total of 88.64% of the interaction sum square. In general, deviation from regression coefficient, AMMI stability value and genotype selection index revealed that Baji, Birmash, Emahoy, IS9302 and Gambella-1107 were relatively stable varieties with optimum grain yield and therefore recommended for further demonstration and popularization in the test locations and areas with similar agro-ecologies.