Studies on conversation repair can reveal the similarities and differences between special populations and normal populations in terms of pragmatic and cognitive abilities. It can provide some references for clinical diagnosis and intervention practices. Therefore, the related studies have attracted the attention of researchers at home and abroad. This paper conducts a systematic review of studies on conversation repair of special populations from Web of Science Core Collection and Scopus by use of bibliometric and content analytic research methods. It is found that foreign studies on conversation repair of special populations show fluctuations in growth, and the empirical studies are in the majority. Based on elicited data, concentrated on patients with cognition impairment, the empirical studies have mainly adopted cross-sectional design to explore whether they have the ability to conduct conversation repair, and to explore whether they have some disorders when they conduct conversation repair. Besides, combined with personal factors and external factors, the empirical studies have explained the mechanism behind conversation repair and analyzed the effectiveness of equipment-supported and non-equipment-supported intervention methods. On the whole, the empirical studies show three trends: (1) the subjects tend to be more diverse; (2) the conversation data tend to be more natural; (3) the research contents tend to deepen constantly. Such results have certain significance for promoting in-depth studies and intervention for special populations in China.