Treatment research typically examines what works for the average individual. In positive psychology, researchers have shown that diverse strategies such as expressing gratitude, savoring experiences, using strengths, increasing optimism, and practicing kindness all demonstrate the potential to boost an individual’s level of well-being. No research, however, aims to help an individual select which of these techniques would most likely benefit him or her. This dissertation addresses this question by creating and validating a system in order to recommend specific positive psychology exercises. I conducted a series of studies to develop and test a recommendation framework for six positive psychology exercises: active-constructive responding, blessings, gratitude visit, life summary, savoring and strengths. In Study 1, 792 participants received up to six positive psychology exercises. After each exercise, participants indicated their preference for each exercise and how often they engaged in it. A factor analysis of these scores revealed three groupings of subjective preferences: active-constructive responding and savoring; blessings and life summary; and gratitude visit and strengths. Individuals who had high preference for an exercise were more likely to complete the exercise. In Study 2, I used these groupings to create a recommendation framework. The sample consisted of 127 undergraduate students who participated in the study over a four-week period. All participants randomly received an initial positive psychology exercise for one week and rated their preference for the exercise. Participants were randomized to either a matched or control group: In the matched group, individuals received a second exercise based on a previously defined matching rule, whereas in the comparison group, individuals received a second exercise by random assignment. Individuals in the matched group preferred the second exercise significantly more and tended to report larger boosts in well-being following the second exercise than those in the control group. I discuss these findings and their implications for adopting idiographic methods to create packages of interventions.