Soil acidity reduces maize (Zea mays L.) yields on about eight million hectares in developing countries. We report on response to selection for tolerance to soil acidity, using an altered version of modified ear‐to‐row (MER) and full‐sib (FS) selection. In the MER selection, 120 half‐sib (HS) families were evaluated under 45 and 80% AI saturation. One to three ears from each of the best 30% of the families were selected each cycle. After 16 cycles of MER selection, FS selection was initiated. Two hundred and fifty FS families were evaluated at five to six acid (ASEs) and one normal soil environments (NSE) the best 25% were selected each cycle. Cycles 2, 4, 6, 8, 10, 12, and 14 of MER and 0, 1, and 2 of FS selection were evaluated in three to nine replications at six ASEs and five NSEs during 1990–1991. Across the 11 environments, gain from selection in yield averaged 40 kg ha−1 cycle−1 (1.49**%) with MER and 250 kg ha−1 cycle−1v (8.10*%) with FS selection. Across the six ASEs, yield improvements of 40 kg ha−1 cycle−1 (1.99**%) with MER and 310 kg ha−1 cycle−1 (13.96**%) with FS selection were obtained. Yield also improved across the five NSEs by 50 kg ha−1 cycle−1 (1.10**%) with MER and 150 kg ha−1 cycle−1 (3.31%) with FS selection. Results indicate that tolerance to soil acidity can be improved with recurrent selection and that the progress will be higher with a system more efficient at reducing the experimental error and genotype ✕ environment interaction.