The population dynamics of busseola fusca (Fuller) was studied from 1997 to 1999 at Halhale and Halhal Begos, located at altitudes of 1960 and 1850 m above sea level respectively in the highlands of Eritrea. Sowing was done in early April in Halhale and late June in Halhal Begos throughout the study period. All agronomic practices recommended for sorghum production in the two areas were followed to ensure good crop stand, except for application of pesticides. A total of 15 generations, nine in Halhale and six in Halhal Begos were observed during the study periods. Busseola fusca had three and two generations per year in Halhale and Halhal Begos, respectively. Peak egg laying at Halhale was in early May, July and early September, while at Halhal Begos it was in July and early September. The study revealed that stage-specific mortalities of B. fusca were caused by infertility, disappearance (predation by ants and earwigs, lack of suitable host, abiotic factors), residuals and parasitoids. The key mortality factor was disappearance of small larvae (35–66 % mortality). The mortality of the other stages was low but had a composite effect on the overall generation mortality. The mortality of small larvae was density-dependent and density-independent at Halhale and Halhal Begos, respectively. The changes in population between generations were due to small larval mortality of B. fusca from one generation to the next. The life table study showed that mortality due to natural factors did not reduce B. fusca population below the damage level. Integrated management strategies are discussed to reduce the damage caused by B. fusca on sorghum.