The Difficulties in Emotion Regulation Scale (DERS) assesses emotion regulation among adults and adolescents, in healthy as well as clinical populations. However, its factor structure is a widely debated issue. Although earlier studies replicated the original six-factor structure, they pointed out psychometrically poor items in the DERS. The issue is further complicated by the findings of recent studies reporting a five- or even a four-factor structure of the DERS. Further, several studies have questioned the validity of the Awareness dimension of DERS and suggested excluding this dimension. Against this backdrop, the present study attempts to examine the factor structure of the Hindi version of the DERS (DERS-H) on 434 healthy participants (303 females, 131 males) of middle socio-economic status. Findings of the exploratory and confirmatory factor analyses (CFA) did not support either of the one-, six- or five-factor solutions. After excluding five psychometrically poor items (poor loadings and higher standardized residuals), the CFA though yielded a satisfactory fit for first-order six-factor model, the second-order model did not fit well. However, after removing the Awareness dimension, the first-order as well as the second-order modified five-factor models provided apt fits to the data. This modified DERS-H (and its five sub-domains) showed satisfactory reliability and validity as well as factor structure invariance across gender and age. The findings suggest that the modified DERS-H is a psychometrically sound tool (with stable factor structure) to measure the second-order construct of emotion regulation difficulties as well as its five first-order subdomains, particularly in the Indian cultural context.