According to the literatures on PIU, a rather large body of researchers has provided empirical evidence to support the idea that the quality of interpersonal relationships may contribute to PIU. But until now, little is known about the mechanisms underlying attachment orientations on PIU. Davis (2001) presented a cognitive-behavioral model of PIU, and thought that the maladaptive cognitions of self and world play critical roles in PIU. Based on the cognitive-behavioral model of PIU, this study presents a unique perspective on the underlying mechanisms of attachment orientations on PIU by focusing on the possible mediating roles of social self-efficacy and loneliness. Using cluster random sampling method, total of 483 undergraduates (250 men, 233 women;) were recruited from three universities in Guangdong province, aged from 17 to 25 years (M = 19.82, SD = .97). Participants completed paper-and-pencil version of the questionnaires of adult attachment (attachment anxiety and attachment avoidance), social self-efficacy, loneliness and PIU (core symptoms and related problems). Participants were informed about the voluntary nature of the study, and obtained presents in exchange for participating. In this study, the hypothetical model was estimated using the maximum likelihood method (ML) and the bootstrap method in the Mplus program (version 7.0). In addition, as recommended by Matsunaga (2008), the current study adopted item parceling strategies (factorial algorithm and internal-consistency approach) to solve the problem that a single latent variable contains too many observation indexes. Because independent samples t tests of gender differences revealed that significant differences between women and men on study variables emerged, the variable gender was controlled in the subsequent mediation modeling analyses. Analyses of the mediation model showed a good fit to the data (χ2 = 382.99, df = 152, χ2 / df = 2.52, p ≤ .001; RMSEA = .056; CFI = .961; TLI = .951; SRMR = .043), and revealed that: (1) The relationship between attachment anxiety and PIU was mediated by loneliness, and the significant mediating roles were composed of: the mediating roles of loneliness for core symptoms (effect size = 25.10％) and for related problems (effect size = 26.27％); (2) Regarding to attachment avoidance dimension, the mediating role of loneliness and the serial mediating role of social self-efficacy via loneliness were found in this linkage, and the significant mediating roles were composed of: (a) the mediating roles of loneliness for core symptoms (effect size = 33.89％) and for related problems (effect size = 33.33％); the serial mediating roles of social self-efficacy via loneliness for core symptoms (effect size = 49.59％) and for related problems (effect size = 48.72％). These results basically supported our theoretical hypotheses that social self-efficacy and loneliness mediated the relationship between attachment orientations and PIU. To our knowledge, this is the first study of its kind, and it expands upon the growing corpus of research relating the relationship between attachment orientations and PIU by understanding the mechanisms leading to attachment anxious and avoidant individuals’ lower social self-efficacy and higher loneliness. These findings have theoretical and applicative value for the prevention and clinical treatments of PIU.