Maize is one of the first crops to produce hybrids using cytoplasmic male sterile lines. The C-type cytoplasmic male sterile (CMS-C) line is vital for hybrid seed production, and the fertility-restoration reaction along with its stability has a direct bearing on its applications. However, fertility restoration mechanism of CMS-C is complex and is still not clear so far. To further explore the factors affecting the fertility restoration of maize CMS, a series of test crosses were carried out by pollinating the isocytoplasm allonuclear CMS-C lines C48-2, Chuangzaosi and C478 with the test lines 18 bai, zi 330, 5022 and the restorer line A619, respectively. Four F2 populations and six double-cross combinations were obtained from the self-cross of fertility restored F1 and pollinating male-sterility-maintained F1 with the male-fertility-restored F1, respectively. Meanwhile, we developed the incomplete diallel-cross combinations using the isocytoplasm allonuclear male sterile lines as maternal parents and their respective maintainer lines 48-2, huangzaosi and 478 as paternal parents. All the F1, F2 and double-cross populations were planted at distinct locations in different years, and the fertility-restoration reaction was scored by field investigation and pollen staining with I2-IK. The results were as follows: 1) The same test line could restore the CMS-C line at a certain genetic background, but failed to restore the CMS-C line at the other genetic backgrounds, suggesting that the genetic background of CMS-C lines plays an important role in the fertility restoration. 2) The fertile-to-sterile segregation ratio of (C48-2×A619) F2 population planted in both Sichuan and Yunnan fited well to 15:1 by the χ 2 test. However, the fertility level of individuals in Yunnan mainly belonged to the 3 and 4grades, but which in Sichuan mainly belonged to the 5 grade, indicating the environmental factors had effect on the fertility-restoration reaction of (C48-2×A619) F2. 3) In our study, 18 bai could not restore C478, and 48-2 could not restore C478, but the fertile and sterile segregated plants were unexpectedly found in their double-cross population [(C478×18 bai) F1S×(C48-2×18 bai) F1F]. The similar case was also observed in the double-cross population [(C48-2 × zi 330) F1S × (C478 × zi 330) F1F]. Therefore, we speculated that there are minor fertility restorer genes not only in the nuclear background of C48-2 and C478, but also in zi 330 and 18 bai, and when these minor genes were gathered by hybridization, they could restore the fertility of C478 and C48-2. This conforms to the restorer genes dose-effect for fertility restoration in the plant CMS system. These results not only contribute to our understanding of the complexity and diversity of CMS-C restoration mechanism, but also provide an important reference for the practical applications about maize CMS-C.