Abstract This research examined the influence of emotional trauma after break-up on self/other working models with the Implicit Association Test in the context of adult attachment. We tried to explore the differences of definite self/other working models on the subjects having and having not experienced emotional trauma after break-up with different attachment styles. We also tried to test and verify Attachment Theory’s hypothesis on self/other working models of different attachment style subjects by means of social cognitive experimental method, such as the Implicit Association Test. The results indicated that:(1) Although the scores of attachment anxiety and attachment avoidance were not significantly different between the experimental group and the control group, the main effects of adult attachment style and grouping existed in IAT effect (indicator of self- positive/other-negative working model) and the interaction was also significant. In addition to avoidant attachment subjects, the IAT effect values of secure attachment subjects and anxious attachment subjects in the experimental group were significantly higher than the IAT effect values of the control group, which indicated that emotional trauma after break-up had prompted the subjects of secure attachment and anxious attachment strengthening the use of ego defense mechanisms, their attitudes and evaluations towards self were all significantly better than their attitudes and opinions towards the ex-lover. However, avoidant attachment subjects showed that they weren’t traumatized by the emotional trauma after break-up, which was consistent with many previous researches. This research also testified that it would exist in a subconscious level. (2) This research identified Attachment Theory’s hypothesis on self/other working models of different attachment style subjects. Specifically, in the control group, the response time of the compatible part and the incompatible part of different attachment style subjects were analyzed with paired-t tests, the results were significant; the IAT effect values of different attachment style subjects were also significant. We found that the response time of the compatible part of anxious attachment subjects was significantly longer than the incompatible part, the response time of the compatible part of avoidant attachment subjects was significantly shorter than the incompatible part, and there was no significant difference between the response time of the compatible part and the incompatible part in secure attachment subjects. The IAT effect values of avoidant attachment subjects were significantly higher than other attachment style subjects.