Socio-emotional interaction is considered a crucial factor that may affect the effectiveness and efficiency of collaborative argumentation and deserves further exploration. This study explored what temporal and sequential patterns of socio-emotional interaction could promote or hinder productive collaborative argumentation by comparing the differences between four high- and four low-performing groups. The 8 target groups were selected from a larger sample of 14 groups according to their scores of groups' argument maps after participating in a 90-min collaborative argumentation activity. Using content analysis method, groups' video recordings were coded and analyzed. The results indicated several differences between the high- and low-performing groups. First, more positive socio-emotional interactions were found in the high-performing groups while more negative socio-emotional interactions were discovered in the low-performing groups. Second, using smiles or laughter to express disagreement (P2) and interrupting others (N1) were more essential factors that may affect argumentation quality in the Chinese context. Third, temporal analysis suggested that in low-performing groups, the sharp rise in negative socio-emotional interactions may be provoked by more cognitive interactions in the initial phase that aimed to persuade opponents. Fourth, sequential analysis showed that repeated negative interactions in low-performing groups may lead to a negative atmosphere and thus, made the increasing tension less likely to be relieved by positive interactions. Whereas in high-performing groups, frequent transitions between positive socio-emotional interactions helped create a positive atmosphere, which could avoid conflict escalation and resolve the potential tension caused by negative interactions. Finally, pedagogical implications were suggested.