Structural equation modeling (SEM) is a family of related statistical techniques that lend themselves to understanding the complex relationships among variables that differ among individuals in the population. SEM techniques have become increasingly popular in the study of personality disorders (PDs) and maladaptive personality traits. The current article takes a critical look at the ways in which SEM techniques have been used in the study of PDs, PD symptoms, and pathological personality traits. By far the most common use of SEM in the study of PDs has been to examine the latent structure of these constructs, with an overwhelming bulk of the evidence in favor of a dimensional, as opposed to categorical, conceptualization. Other common uses of SEM in this area are factor models that examine the joint multivariate space of PDs, maladaptive personality traits, and psychopathology. Relatively underused, however, are observed or latent variable path models. We review the strengths and weaknesses of the work done to date, focusing on ways that these SEM studies have been either theoretically and/or statistically sound. Finally, we offer suggestions for future research examining PDs with SEM techniques.