A modified version of the dual-route cascaded (DRC) model argued that the deaf adults access phonology of printed words indirectly through orthographic, such as fingerspelling, or speech-reading routes. Precisely, researches from alphabetic language found the congenital deaf could recode printed words based on the handshape-letter correspondence rule (i.e. visual phoneme awareness). For the hearing, a phoneme position effect was evident in both English (the first phoneme advantage) and Chinese (the first consonant/Shengmu advantage) written words. In line with which, a Shengmu advantage in deaf was found in single Chinese character. However, there has been no evidence so far proving a Shengmu advantage as the same in Chinese disyllable words, so as its causes. A phoneme identification experiment was conducted to investigate whether there would be a similar phoneme position effect (i.e. consonant/Shengmu recognition advantage) in Chinese disyllable words of the deaf, which would be helpful to reveal the mechanism of deaf adults’ phoneme encoding. 64 high-school students with severe hearing loss (>90dB) participated the experiment but with 4 students missed. After a practice session, participants entered the task (192 Chinese disyllable words in total) in which they were asked to judge whether a disyllable word contain the consonant/vowel phoneme or not. No feedback was given during the task. The present study consisted of two parts. Specifically, a two-factor within group design of 2(phoneme type: consonant/Shengmu, vowel/Yunmu) *2(phonological code type: fingerspelling, letter) was conducted at the first to investigate the initial consonant/Shengmu recognition advantage and the influence of phonological codes. On the basis of which, a two-factor within group design of 2 (phoneme type: consonant/Shengmu, vowel/Yunmu) *2(morpheme position: beginning character, end character) was conducted to further explore the influence of morpheme position on the consonant/Shengmu recognition advantage in Chinese disyllable words. The results indicated that: 1) the performance of consonants was more accurate and faster than vowels in disyllable words, revealing a consonant recognition advantage of the deaf. 2) fingerspelling and visual letter were both used by the deaf to encode phonology in printed words while fingerspelling code was less effective, in addition, the initial consonant recognition advantage was independent of two types of phonological codes. 3) a significant morpheme position effect was found in disyllable words, i.e. the phoneme recognition accuracy was higher and the speed faster for the beginning character than the end character in Chinese compound words; further, the performance for the consonant/Shengmu of first character in compound word was significantly better than all other phonemes, both indicating that for deaf students, not only syllables in compound word but also phonemes within a syllable were both serially, from left to right, encoded. In conclude, the findings thus indicated that there was a significant Shengmu recognition advantage in disyllable word of the deaf, jointly influenced by fingerspelling and the serial phoneme encoding. The findings of the present study will not only improve our understanding of the phenetic encoding mechanism in printed Chinese words, but also provide insights to the language instruction for the deaf.