(1) Background: Currently, the scientific evidence on the benefits of assisted therapy with dogs in dementia is not clear. In this study, we want to evaluate such benefits through a randomized controlled clinical trial in multiple centers across the country. (2) Methods: The participants were people over 65 years old with dementia, residing in senior centers in Spain (n = 334). The experimental group underwent assisted therapy with dogs based on the Comprehensive Cognitive Activation Program in Dementia, for 8 months, with weekly sessions of 45 min. Data were collected at the commencement, middle, and end of the program, to evaluate the aspects using the Mini-Examination Cognitive, the modified Bartell Index, the Cornell Scale for Depression in Dementia and the Neuropsychiatric Inventory. (3) Results: The results show significant improvements in the experimental group versus the control group in the affective (T1 = p 0.000; T2 = p 0.000) and behavioral (T1 = p 0.005; T2 = p 0.000) aspects, with the affective aspect displaying greater progress in participants with additional depressive (p = 0.022) or anxiety (p = 0.000) disorders, shorter institutionalization periods (r = −0.222, p = 0.004), and those undergoing complementary psychotherapy (p = 0.033) or alternative therapy (p = 0.011). (4) Conclusions: Dog therapy is effective in improving the affective and behavioral aspects of institutionalized patients with dementia.