Heart rate variability (HRV) mirrors autonomic nervous system activities and might serve as a parameter to monitor health status in older adults. However, it is currently unknown which functional health measures, including cognitive, physical, and gait performance parameters, are most strongly related to HRV indices. This knowledge would enable implementing HRV assessments into health monitoring routines and training planning for older adults. Simultaneous cognitive–motor and exergame training may be effective to improve HRV indices but has not been investigated yet. Eighty-nine healthy older adults (≥70 years of age) were randomized into three groups: (1) virtual reality video game dancing, i.e., exergaming (DANCE); (2) treadmill walking with simultaneous verbal memory training (MEMORY); or (3) treadmill walking only (PHYS). Strength and balance exercises complemented each program. Over 6 months, two weekly 1-h training sessions were performed. HRV indices (standard deviation of N–N intervals, SDNN; root mean square of successive R–R interval differences, RMSSD; and absolute power of high-frequency band (0.15–0.4 Hz), HF power) and various measures of cognitive, physical, and gait performance were assessed at baseline and after 3 months and 6 months. Multiple linear regression analyses with planned comparisons were calculated. At baseline, 8–12% of HRV variance was significantly explained by cognitive executive functions and leg strength (inversely related). Verbal long-term memory, aerobic and functional fitness, and gait performance did not contribute to the model (SDNN:
R
2
= 0.082,
p
= 0.016; RMSSD:
R
2
= 0.121,
p
= 0.013; HF power:
R
2
= 0.119,
p
= 0.015). After 6 months, DANCE improved HRV indices, while MEMORY and PHYS did not (time × intervention interactions: first-contrast DANCE/MEMORY vs. PHYS: SDNN
p
= 0.014 one-tailed, Δ
R
2
= 0.020 and RMSSD
p
= 0.052 one-tailed (trend), Δ
R
2
= 0.007; second-contrast DANCE vs. MEMORY: SDNN
p
= 0.002 one-tailed, Δ
R
2
= 0.035, RMSSD
p
= 0.017 one-tailed, Δ
R
2
= 0.012, and HF power
p
= 0.011 one-tailed, Δ
R
2
= 0.013). We conclude that mainly cognitive executive functions are associated with HRV indices and that exergame training improves global and parasympathetic autonomic nervous system activities in older adults. Periodic assessments of HRV in older citizens could be particularly beneficial to monitor cognitive health and provide indications for preventative exercise measures.