A variety improvement strategy was adopted through a coordinated, collaborative and complementary approach to breed for resistance to foliar diseases of maize in Kenya. The objective was to breed for resistance to priority pests and diseases identified as key constraints in maize-growing ecosystems. This is the only economic solution towards achieving high maize yields in disease-prone environments. The main diseases are turcicum leaf blight, gray leaf spot, common rust, head smut and maize streak virus (MSV). Priority pests are stem borers and Striga, or witchweed. MSV is one of the major biotic constraints to increasing maize yields in Kenyan ecosystems II, III and IV. Eight breeding populations of maize were constituted in 1999 by crossing elite x elite lines. The populations were MU002, MU007, MU015, MU016, DC17, DC31, DC96 and the DC Complex population. The populations were initiated primarily for line recycling and development of pure lines to be used as parents of hybrids. At the Kenya Agricultural Research Institute station at Muguga, advance at every stage of generation within and between row selection was done for adaptability and resistance to MSV and common rust. Screening for MSV resistance in each population was done at the S1 generation with a final screen at the S4:6 stage. The results of the MSV screen showed that each population behaved differently in the number of lines identified as highly resistant to MSV. At the S4:6 level, 85% of the lines of MU007 showed a highly resistant score (< 1.5); MU016, 89%; MU002 and MU015 populations both had 14% of lines with high resistance to MSV and 23% of lines with moderate resistance. Thirty-seven lines selected from two populations, MU015 and MU002, were used in population improvement. One hundred two three-way cross hybrids were generated by crossing each line to two single-cross testers, CML395/CML444 and CML312/CML442. Crosses were evaluated for performance in Kakamega and Bukura to identify the best-performing three-way cross hybrids. Twenty hybrids were identified as best performers in terms of grain yield and resistance to MSV. The best yield of the best hybrid, MU03-032, was 12.62 t ha–1, which was significantly better than the check mean of 10.78 t ha–1 (p < 0.05). These results indicate potential three-way cross hybrids that could be directly deployed for use by farmers in MSV-prone environments.