Speaking involves stages of conceptual preparation, lemma selection, word-form encoding and articulation. Furthermore, process of word-form encoding can be divided into morphological encoding process, phonological encoding process and phonetic encoding. What is the function unit at the stage of word-form encoding remains a controversial issue in speech production theories. The present study investigated syllable and segments effects at the stages of phonological encoding, phonetic encoding, and articulation in Mandarin spoken word production. Using Picture-Word Interference (PWI) Paradigm, we compared the effects generated in immediately naming (experiment 1), delayed naming (experiment 2), and delayed naming combined with articulation suppression (experiment 3) tasks. Eighteen black and white line drawings were applied as stimuli, and their names were monosyllabic words. Each target picture was paired with four distractor words: A CVC-related (C: Consonant, V: Vowel) distractor word was chosen that shared a syllable which always differed in tone with the picture name (i.e.,羊 /yang2/ as target name -央/yang1/ as distractor word). A CV-related distractor word was chosen that shared the onset consonant and the core vowel with the picture name (i.e., 羊/yang2/-药/yao4/). A VC-related distractor was chosen that shared the rhymes with the picture name (i.e., 羊/yang2/-让/rang4/). An unrelated distractor was selected that stood in no obvious semantic, phonological or orthographic relation with the picture name. We found syllable and segments facilitation effects in immediate naming, whereas syllable and segments inhibition effects in a delayed naming and a combination task of delayed naming and articulation suppression. An immediate naming involves stages of phonological encoding, phonetic encoding, and articulation, a delayed naming involves articulation only, while a combination task of delayed naming and articulation suppression involves phonetic encoding and articulation processes. By comparing these effects among three tasks, we suggest that syllable and segments facilitation effects localized at the stage of phonological encoding, whereas syllable and segments inhibition effects localized at the stage of phonetic encoding and (or) articulation. These findings indicated that syllable plays a more important role in phonological encoding whereas segments play their roles in phonetic encoding and articulation for motor programming. Our findings provide support for Proximate Unit Principle and the assumption of independence of premotor- (phonological encoding) and motor stages (phonetic encoding and articulation).