The main purpose of this research was to use SEM model to study the relationships among adult attachment, social self-efficacy, distress self-disclosure, loneliness and depression of the college students who had romantic experience. The participants of the study were 805 Taiwan college students who had romance, and were divided into two groups. Data from the participants in group one were used for instrument and model development, and data from the participants in group two were used for model validation. The researcher found the model was well supported by observed data and the stability of the model was confirmed. The main results were as follows: (a) Attachment avoidance and attachment anxiety affect social self-efficacy, which in turn affects loneliness and further indirectly affect depression through loneliness; (b) Attachment avoidance affects distress self-disclosure, which in turn affects social self-efficacy and sequentially affects loneliness sand depression; (c) Attachment anxiety directly affects depression; (d) Attachment anxiety indirectly affects depression throughloneliness; (e) Attachment anxiety is a stronger predictor for depression thanattachment avoidance; and (f) Distressself-disclosure indirectly affects loneliness through social self-efficacy.