Aim The study assessed the effects of different weekly training frequencies performed over a 12 week exercise programme (strength, endurance, balance) on autonomic cardiac activity, blood pressure and cardiorespiratory fitness (CRF) in elderly individuals. Methods Fifty-eight individuals participated in the study: 2TG (N = 24, 71.1 ± 6.4 yrs; 19 females, 5 males) performing two, 60 min sessions/week; and 3TG (N = 34, 72.3 ± 7.9 yrs; 25 females, 9 males) performing three, 60 min sessions/week. Time domain and spectral analysis of heart rate variability (HRV) quantified autonomic cardiac regulation. Results Natural logarithm (Ln) transformation was applied to all HRV parameters. There were significant reductions in total power (Ln TP) (p = 0.006), low frequency (Ln LF) (p = 0.013), high frequency (Ln HF) (p = 0.013) and root mean square of successive differences (Ln rMSSD) (p = 0.014) post training in 3TG after intervention. Diastolic BP (DBP) decreased significantly in both groups (2TG: P < 0.001; 3TG: P < 0.001). Both groups showed significant improvements in six-minute walk distance (2TG: P = 0.003, 3TG: P = 0.001). However, there were significant HRV differences between 2TG and 3TG for Ln TP (P = 0.018), Ln LF (P = 0.049), Ln HF (P = 0.039) and Ln rMSSD (P = 0.049). Conclusions A combined exercise programme resulted in improved DBP and CRF irrespective of training two or three 60 min sessions/week. However, training three, sessions/week induced negative health-related changes in autonomic cardiac activity through reducing HRV parasympathetic function, while HRV was maintained in the group training twice a week.