Background: About three-quarters of people with dementia live in their own homes, with help from family members and/or other unpaid carers, such as friends or neighbors. Often, unpaid carers themselves experience negative consequences, such as stress, burden, and symptoms of depression or anxiety. Research has shown that these consequences can be alleviated by psychosocial and psychological interventions. Moreover, there are indications that those interventions can be effective when offered online. Objective: This paper describes the protocol of a randomized controlled trial (RCT) that will take place in the Netherlands to evaluate the effectiveness of iSupport, a minimally guided, internet-based intervention to improve carers' mental health and coping resources. Methods: A superiority two-arm RCT comparing the effects of the online support program with a waiting list control condition will be carried out in the Netherlands. The iSupport intervention was developed by the World Health Organization and is based on cognitive behavioral therapy principles. It has five main themes divided into 23 lessons. Carers can pick and choose which lessons they want to complete. We aim to recruit 200 unpaid carers. The experimental group (n=100) will be provided with access to the intervention for 3 months following randomization; those in the waiting list control group (n=100) will be granted access to the intervention after 3 months. Assessments will be conducted at baseline (T0), 3 months after baseline (post intervention, T1), and 6 months after baseline (follow-up, T2). The primary outcome is perceived stress, measured by the Perceived Stress Scale. Secondary outcomes are symptoms of depression and anxiety, caregiver burden, sense of competence, self-efficacy, mastery, and carers' attitudes toward dementia and their person-centered approach (ie, to what extent carers tailor the provided care to the interest, needs, and history of the person with dementia). Results: Recruitment for the trial started in January 2019. As of July 2019, we have enrolled 120 participants. Data collection is expected to be completed by March 2020. Once all the data have been collected, we will conduct the data analyses between April and May 2020. We aim to publish our results in a manuscript by June 2020. Conclusions: Online interventions have shown promising results in improving the mental health of carers of people with dementia. Additionally, online interventions may overcome accessibility barriers. If successful, this intervention will have important potential for implementation as a public health intervention, since costs and support by trained staff are minimal.