Guinea-race sorghums, originating in western Africa, possess a suite of adaptive traits that make them the predominant sorghums in that region and vital to food security. However, Guinea-race sorghums give limited yield response to intensified production conditions. A promising approach to increase their productivity while retaining the required adaptive and quality characteristics is Guinea-race hybrids. This opportunity was not previously explored due to lack of male-sterile Guinea-race female parents. Research was thus initiated to establish the basis for Guinea-race hybrid development and test this concept. Specific objectives were to 1) create the first Guinea-race parental lines, 2) produce the first hybrids and establish methods for breeding Guinea-race hybrids, and 3) determine the degree and pattern of heterosis in hybrids based on Guinea-race parents to guide breeding efforts effectively. The first series of male-sterile parents (A lines) to be developed includes 16 landraces and 7 interracial lines. The landrace parents are tall, photoperiod sensitive, and possess typical Guinea grain and panicle architecture. These new A lines provide large diversity for most agronomic traits; spanning the range of grain size (1.1 to 3.5 g per 100 seeds), glume form (margaritiferum to conspicuum in the Snowden classification), panicle length (30 to 60 cm) and plant height (3 to 5 m) observed in the Guinea race. The interracial A lines, derived by crossing a Guinea landrace (Bimbiri Soumale) with Caudatum varieties, are dwarf, basically photoperiod insensitive and possess relatively small grain of typical Guinea form. The first experimental hybrids were produced in 2004 on both the interracial A lines (164 hybrids) and landrace A lines (159). Multilocational nurseries of hybrids and their parents were established to observe fertility restoration and initial agronomic characteristics. Collaborative regional hybrid yield trials were initiated for the two target zones: the northern Sudanian zone (Cinzana-IER, Saria-INERA, Bambey-ISRA) and the southern Sudanian zone (Sotuba-IER, Farakoba-INERA, Bengou-INRAN, Kolda-ISRA, Samanko-ICRISAT). A heterosis study was conducted over 2 years in Burkina Faso, Mali and Niger with hybrids produced using 2 testers and 17 male parents representing Guinea sorghums from western, eastern, southern Africa and Asia. Significant positive better-parent heterosis was observed in all environments and with parents from all regions of origin. The average grain-yield superiority over the better parent was 0.41 t ha–1 for the tall female hybrids and 0.85 t ha–1 for the short female hybrids. The best hybrid showed an average yield superiority over the best local Guinea-race checks of 70 to 90%, based on yield advantages of 1.2 to 1.8 t ha–1. These results prove that large, agronomically meaningful yield advantages can be obtained with Guinea-race hybrids. Our diverse set of Guinea-race A lines now permits identification of the most promising parental combinations, and opens the way to identifying the first commercial sorghum hybrids for West Africa.