Abstract Objective The purpose of this study was to investigate the effects of adding virtual reality (VR) to maintenance cardiac rehabilitation (CR); it was hypothesized VR would increase adherence, motivation, and engagement. Methods This study was a randomized, 1:1 concealed-allocation, single-blinded, 2 parallel-arm, crossover trial. Blinded assessments were undertaken at baseline (midprogram), 12 weeks, and 24 weeks after baseline. The setting was a single CR program of unlimited duration in Brazil. Participants were patients with cardiovascular diseases or risk factors who had been in the program for 3 months or longer. The CR program consisted of 3 supervised exercise sessions per week. In the VR arm, participants had 1 VR session of the 3 per week during the initial 12 weeks of the trial; this was withdrawn the subsequent 12 weeks. Measures were program adherence (% of 3 sessions/week over 12 weeks, ascertained in all participants), motivation (Behavioral Regulation in Exercise Questionnaire 3), and engagement (User Engagement Scale, adapted; vigor, dedication, and absorption subscales); all 3 were primary outcomes. Results Sixty-one (83.6%) patients were randomly assigned (n = 30 to CR + VR); 54 (88.5%) were retained at 12 and 24 weeks. At baseline, participants had been in CR on average 7 years and had high engagement and motivation. CR + VR resulted in a significant increase in adherence at 12 weeks (baseline = 72.87%; 12 weeks = 82.80%), with significant reductions at 24 weeks when VR was withdrawn (65.48%); in the usual CR care arm, there were no changes over time. There was a significant effect for arm, with significantly higher adherence in the CR + VR arm than usual CR at 12 weeks (73.51%). Motivation decreased significantly from baseline to 12 weeks (4.32 SD = 0.37 vs 4.02 SD = 0.76) and significantly increased from 12 to 24 weeks in the CR + VR arm (4.37 SD = 0.36). Absorption was significantly lower at 12 weeks in the CR + VR arm (6.79 SD = 0.37 vs 6.20 SD = 1.01). Conclusion Although VR increased program adherence, interspersing it with usual CR sessions actually decreased patient motivation and absorption. Impact Supplementing a maintenance CR program with VR using “exergames” resulted in significantly greater adherence (8% increase or 3 of 36 sessions), and this was quite a robust effect given it was extinguished with the removal of VR. However, contrary to the hypotheses, offering 1 session of VR per week and 2 of usual CR exercise was related to lower motivation and absorption, which has implications for how clinicians design programs for this patient population.