With the development of the Behavioral Activation for Depression Scale (BADS) by Kanter, et al. [1], different behavioral aspects of depression like activation, rumination or avoidance, and functional impairment can be measured. The long and the short versions of the BADS, however, show differences in the quality of psychometric properties. We wanted to validate the short and long forms of the BADS. We therefore evaluated the factor structure, the psychometric properties, and the construct validity of the long version and the subscales in a clinically depressed German sample (n=258) in study 1. In study 2, we explored the factor structure and the psychometric properties of the short version of the BADS in a subsyndromal sample with elevated depressive symptoms (n=406). In study 1, our results replicated the four factor solution of the BADS-25 original version and showed good psychometric properties. However, with regard to the BADS-9 only the French factor solution of the short BADS-9 demonstrated acceptable fit with low internal consistency of the subscale Avoidance. Thus, only the total score of the short form can be recommended.