The Adult ADHD Self-Report Scale (ASRS; Kessler et al., 2005) is one of the most extensively used scales to detect attention-deficit hyperactivity disorder (ADHD) in adults. The aim of this work is to analyze the psychometric properties of the 18 ASRS items in people with substance use disorders (SUDs). Furthermore, we aimed to (a) confirm or, if necessary, modify the dichotomization criteria of the items proposed by the authors, and (b) identify the most informative items for a screening version or, when applicable, confirm the use of the 6 items that comprise the initially proposed short version. The ASRS was completed for 170 patients with SUD at the Provincial Unit for Drug Dependence of Huelva, Spain, aged 16 to 78 years. Two Rasch models—the dichotomous Rasch model and the Rating Scale Model (RSM) for polytomous items—were used in the psychometric analysis. The ASRS items fitted the RSM adequately, but the locations of the items along the underlying construct led us to propose new criteria of dichotomization. After analyzing the information function of dichotomized items, we identified 6 items that should integrate a new screening scale. Our dichotomization proposal is different from the original one and takes into account the different weights of the items. The selected screening version showed better metric properties than the other analyzed versions. Future research should test our proposal by using external criteria and to obtain evidences for other populations, cultures, or patient profiles.