According to attachment theory, the main goal of the caregiving behavioral system is to relieve others’ distress and promote their health and welfare. This is accomplished through a set of caregiving behaviors that are the primary strategy of this system. However, some individuals develop nonoptimal or secondary strategies (hyperactivation or deactivation). The Caregiving System Scale (CSS) is a self-report measure that assesses these nonoptimal caregiving strategies. This study examined the factor structure of the Portuguese version of the CSS using a bifactor model and a 2-factor correlated model; it also examined validity evidence for CSS scores in relation to other relevant variables. The sample comprised 417 women from the general population who completed the CSS, with a subsample of 124 women completing additional measures of attachment, mental representations of caregiving, compassion, and difficulties in emotional regulation. The bifactor model showed the best fit to the data, supporting an orthogonal and reliable 2-factor structure (Hyperactivation and Deactivation). This model also suggested that the Hyperactivation scale is multidimensional. With regard to the validity evidence for the CSS, the results showed that individuals engaged in volunteering activities presented lower levels of deactivation than those who were not engaged in these activities. The CSS scores also correlated as expected with the other measures. In conclusion, this study supports the utilization of Deactivation and Hyperactivation subscales as 2 statistically distinct constructs and demonstrates that the Portuguese version is an adequate measure of nonoptimal caregiving strategies.