The purpose of the present study was to examine the factorial and convergent validity of the German translation of the Exercise Identity Scale (EIS; Anderson & Cychosz, 1994). We introduce a bifactor model as a possible alternative to the previously discussed one- and two-factor models. Using an online study with 530 participants, our results suggest that the EIS should be considered a unidimensional measure. For this one-factor model, temporal invariance (time interval: 2 weeks) and invariance between men and women could be established. Relations to similar constructs were in the expected directions. We provide a psychometrically sound German version of the EIS, which demonstrated factorial and convergent validity in this study. The use of the scale in German samples is likely to advance research on exercise adoption and maintenance by integrating the exercise identity construct.