In modern Chinese, homomorphemic monosyllabic and disyllabic words are one of the very special groups of word clusters which are easily confused by Chinese as a second language learners. Previous studies have not comprehensively examined this confusion phenomenon, and there is a lack of relevant literature involving psychological analyses. Firstly, this article comprehensively surveys the Chinese Intermediate Language Corpus and identifies 27 groups of highly confusable words. Secondly, this article analyses the distribution characteristics of the homomorphemic monosyllabic and disyllabic words, finding strong commonality in the distribution of learners’ native language backgrounds, with overwhelming bidirectional misuse and close lexical-semantic relations. It also finds that verbs are mainly misused in stylistic collocations, and nouns are frequently misused in both prosodic and stylistic collocations, while adjectives are often misused in semantics. Besides, the article analyses the generative mechanism from the psychological perspective of Chinese as the second language learners. It argues that L1-lemma mediation and the lack of prosodic and stylistic information of words in the psychological lexicon of leaners are the main reasons for the confusion. This article argues that the semantic confusion of words is difficult to overcome, while prosodic and stylistic confusion are even more difficult to overcome. Last but not the least, the article offers targeted pedagogical suggestions.