Using the Pearlin stress process model, the present study aimed to test if there was an association between caregiving intensity and caregiver burden, to analyze what type of association existed, and to test if different indicators of social support moderated such association among caregivers of people with dementia. Data from the baseline assessment of the Resources for Enhancing Alzheimer's Caregiver Health (REACH II) (N = 637) were used. Caregiver burden (12-item Zarit caregiver burden scale), caregiving intensity (caregiving hours), and social support (Lubben social network, received support, satisfaction with support, and negative interactions) were the main measurements. Separate multivariate regression models were conducted with Stata 16. The results showed that the relationship between caregiving hours and caregiver burden was a nonlinear inversed U shape after controlling all of the socio-demographic variables. Further analyses showed that when caregiving hours reached 14 hours per day, the levels of burden were the highest. In addition, received support, satisfaction with support, and social network significantly buffered the relationship between caregiving hours and caregiver burden when they were examined separately. However, only social network played a significant buffering role when examining the four social support indicators simultaneously. These findings suggest the need for programs and practices that emphasize the importance of identifying, gaining, and strengthening positive aspect of social support, especially in how to broaden a caregiver's social network while caring for a family member with dementia.