AbstractGeneral combining ability (GCA) referencing additive genetic effects and specific combining ability (SCA) referencing nonadditive effects have been applied for decades in maize (Zea mays L.) breeding and other crops to produce greater genetic gains in crop breeding. This study examined the magnitude of GCA and SCA for cane yield and evaluated potential for application in sugarcane (Saccharum spp.) breeding. Data were collected from unselected first‐stage breeding populations sourced from the irrigated, midlands (high‐altitude), and coastal breeding programs of the South African Sugarcane Research Institute. Cane yield data calculated from number of stalks, stalk height, and stalk diameter was analyzed using best linear unbiased prediction (BLUP) analysis where female and male effects (GCA) and female × male interaction effects (SCA) were defined as random effects. There were significant (P < .05) female but nonsignificant male effects. Female × male interaction was highly significant (P < .001) and larger than both female and male effects, indicating higher SCA than GCA. Specific combining ability variance was 24% (midlands), 31% (coastal), and 57% (irrigated) higher than GCA. Genotypes with low GCA produced higher SCA. Crosses from parents originating from different breeding backgrounds produced higher SCA than those from the same backgrounds. Crosses with significant SCA produced 24–42% greater cane yield than those with nonsignificant SCA. Using GCA to develop parents with high breeding values and SCA to evaluate parent cross combinations will maximize exploitation of heterosis in sugarcane breeding. The results suggest the possibility to develop heterotic groups that are diverse and complementary as a strategy to optimize utilization of SCA.