Estimated autonomic nerve system (ANS) through heart rate variability (HRV) is a well-known practical method to investigate risk factor of cardiovascular disease (CVD) such as sudden cardiac death. Frequency range of HRV directly represents ANS balance, but one of power spectrum (PS) analysis methods should be determined depending on non-stationary factor of R-R interval or recording time. Although autoregressive model (AR) and fast Fourier transform (FFT) theoretically derive similar values; however, two methods cannot be completely replaced in a variety of clinical situations, in particular exercise. PURPOSE: This study aimed at investigating difference, correlation, and interchangeability of HRV index obtained from AR and FFT based on R-R interval at resting and recovery after Wingate test. METHODS: Sixty-seven healthy males without cardiovascular and orthopedic diseases participated. R-R intervals obtained using Polar RS800CX at rest and recovery of Wingate test. Each index was computed from AR and FFT by Kubios. In statistical analysis, Wilcoxon signed-rank used to compare differences among measured variables. Spearman correlation coefficients were calculated. Intra-class correlation coefficients (ICC) yielded to investigate internal reliability. In addition, Bland & Altman plot used to assess limits of agreement for estimated interchangeability. RESULTS: There was no significant difference at the entire periods except for LF/HF ratio (p < 0.05) at recovery. LF, HF, and LF/HF ratio of AR and FFT at resting and recovery showed strong correlation (p < 0.001). Also, high internal consistency in frequency index (p < 0.001) was shown as results of ICC analysis (LF, 0.952; HF, 0.987; LF/HF, 0.923), but Bland & Altman analysis indicated deviated shape of scatter plot and existence of significant bias within 95% confidence interval upper and lower of agreement which represents obvious bias between AR and FFT. CONCLUSION: Although AR and FFT had significant correlation and high internal consistency, they could not be replaced with one another for interpreting HRV in monitor of ANS of Wingate test.