Reading is not only an essential approach to get information, but also a important way to obtain new knowledge for children. The developmental dyslexia is a kind of specific learning disabilities. What is the underlying cause? A lot of researchers have dedicated to this issue. In western alphabetic language system, phonological awareness is recognized as the core deficit and the best predictor of reading. However, in recent years, a number of studies further explore the cognitive mechanism of the developmental dyslexia from the new perspective of working memory. The graphemes – phonemes binding deficits processing has been the focus of developmental dyslexia research during the last decades. Feature binding is an encoded way to store working memory in cognitive psychology, that is to say, people regard the various and complex features as a whole to stimulate memory. Based on the above theory, the basic process of reading is a coding processing that making grapheme – phoneme binding as a unit. Previous study tested the predicted role of visual acuity, phonological awareness and grapheme – sound binding in reading level ability of 8-year-old child who come from China (Hong Kong and Taiwan) and the United Kingdom. The results show that grapheme – phoneme binding is the strongest predictor of Chinese reading. To further explore how grapheme - phoneme binding deficit paradigm may affect children’s developmental dyslexia, cross-channel integration processing is a new trend to explore the fundamental cognitive processing of developmental dyslexia in China. Therefore, there are three kind of methods were used to examine the graphemes – phonemes binding deficit in different language system. The first one is paired- association learning (PAL), were widely used to distinguish developmental dyslexia children and the normal children. And then the second one is change detection task (CDT), which is different from PAL and is more effective to improve the grapheme - phoneme binding deficit from the perspective of feature binding in working memory. Last but not least, a cued recall task (CRT) was usually combine with change detection task, and provide the new clue to make up with grapheme - phoneme binding deficit. In this review, it has also been revealed that the nature of graphemes – phonemes binding processing of developmental dyslexia between non-dyslexia group and dyslexia group by the difference in brain neural mechanism. In alphabetic language system, a lot of functional magnetic resonance imaging studies shown that, the superior temporal sulcus (STS) and the superior temporal gyrus (STG) is the key brain neural mechanism for graphemes – phonemes binding deficit adults group. However, the graphemes – phonemes binding research in Chinese developmental dyslexia still remain in cognitive behavioral level, and disagreements in selecting criteria of subjects and in experimental paradigms. Further research in this area should address the specificity of Chinese dyslexia, and expect to investigate the cognitive and neural mechanism of the graphemes – phonemes binding processing by the empirical research.