Engaging persons with dementia in meaningful activities supports well-being; however, care staff are challenged to implement age- and ability-appropriate activities in a group setting. We compared a randomly assigned treatment group, who received horticultural therapy-based (HT-based) programming to a comparison group, who engaged in traditional activities (TA) programming, on engagement and affect. Horticultural therapy-based programming was implemented twice weekly at 4 treatment sites for 6 weeks, while regular TA were observed at comparison sites. Results revealed no differences between groups on affective domains. Levels of adaptive behavior differed between the groups, with the treatment group demonstrating higher levels of active, passive, and other engagement and the comparison group demonstrating higher levels of self-engagement. Our results highlight the value of HT-based programs and the importance of simultaneously capturing participants' affective and behavioral responses. Theoretical and practical considerations about the facilitation of and context in which the programming occurs are discussed.