DNA methylation in the form of cytosine methylation is proposed as an evolutionary event, which contributes to genome evolution and plays an important role in maintaining genome integrity and controlling dynamics of gene activity. Hybridization and polyploidization play significant roles in the evolution of higher plants. Compared with animals, cytosine methylation is more abundant in plants, particularly at genomic regions containing transposons and their derivatives. Accumulated evidence had suggested that DNA methylation plays important roles in normal plant development. In different stages of plant growth, the changes in DNA methylation level play an important role in response to variations of its heredity and the environment. Notwithstanding these interesting findings, the causing factors for, and generality of, methylation dynamics in plants, particularly with hybrid formation and trans-generational heritability, remained largely vague. However, there is a paucity of evidence to support any direct link between the epigenetic phenomena of cytosine methylation alteration and gene silencing following hybridization and polyploidization. In this study, MSAP (methylation-sensitive amplified fragment length polymorphism) was used in this study to detect the DNA methylation patterns in the 5′-CCGG sites of two cotton hybrids derived from CRI-12 and their parents for understanding developmental stability and inheritance of cytosine methylation. It was found MSAP ratios, which were the ratios of MSAP type in the two cotton hybrids were 12.41–20.05%, cytosine methylation profiles were variable, from increase to decrease during plant growth and development. Full methylation of internal cytosine (6.90–11.47%) was the dominant in two cotton hybrids. Meanwhile, the MSAP profiles enable the monitoring of inheritance or variation of parental methylation patterns in hybrid progenies. It was found that a great majority (from 96.61% to 98.86%, depending on crosses) of the methylation profiles in cotton inbred lines transmitted to the inter-strain hybrids; however, from 1.14% to 3.39% of the profiles in the hybrids exhibited variation from the expected parental additivity. Both inherited and altered methylation profiles can be divided into distinct groups, and their frequencies are variable among the cross-combinations, and during plant growth and development. Bands in hybrids that appeared in both groups digested by Hpa II-EcoR I and Msp I-EcoR I, which were inherited from either or both of the maternal and paternal parent—this type apparently comprised the greatest majority, as all monomorphic bands belong to this type. In addition, sequencing of differentially methylated fragments and subsequent homology analysis of isolated bands that showed variation in hybrids indicated that diverse sequences were involved, including known-function cellular genes and mobile elements, Such as leucine-rich repeat family protein, PDR-like ABC-transporter, putative oligopeptide transporter, GTP-binding protein, similar to pathogenesis-related protein, DOMON domain-containing protein, putative adenosine phosphosulfate kinase, putative protein, RNA-directed DNA polymerase. The remaining 14 bands showed no homology to the database sequences. These results clearly demonstrated the power of the MSAP technique for large-scale DNA methylation detection in the cotton genome, and the complexity of DNA methylation change during plant growth and development. The different methylation levels may be induced by interspecific hybridization between two cotton hybrids, and indicated a direct relationship between cytosine methylation alteration and gene expression variation.