The present study investigated the representation and processing of segmental and tonal information in visual Chinese word recognition in native and non-native Chinese readers. Two experiments using homophone judgement paradigm were conducted. When judging two Chinese characters (Experiment 1), both groups showed difficulties when the segmental but not tonal information was shared (i.e., S+T- condition), with stronger interference for non-native readers. When judging a Pinyin and a Chinese character (Experiment 2), non-native readers' S+T- disadvantage was significantly reduced. Finally, when their L1s share the Latin alphabets with Pinyin, the participants showed stronger segmental reliance compared with the native readers. These findings suggest that while both native and non-native readers rely more on segmental information in Chinese visual word recognition, this reliance is stronger among non-native readers. Furthermore, the alphabetic nature of the Pinyin orthography may help non-native readers' phonological representation and processing in a nonalphabetic writing system.