To achieve efficient reading comprehension, Chinese as a second language (L2) learners adopt various component skills and knowledge, including the aspects of vocabulary, grammar, and coherence. However, there is no consensus regarding the relative predictive powers of these three areas in L2 reading comprehension. Instead of employing direct measurement of student reading performance, this study aims to explore this issue by analyzing teacher perception data. Using a Likert-scale questionnaire, 95 experienced teachers rated L2 Chinese learners' reading comprehension, subskills, and sub-knowledge of reading. The results of structural equation modeling reveal that the three areas of subskills and sub-knowledge (i.e., "Vocabulary," "Grammar," and "Coherence") all positively and significantly predicted reading comprehension. "Grammar" exhibits the strongest predictive power, followed by "Coherence," and "Vocabulary." This study enhances our understanding of the relative contributions of subskills and sub-knowledge to L2 Chinese reading. Implications for L2 Chinese instruction are provided.