Acute hypoxia activates several autonomic mechanisms, mainly in cardiovascular system and respiratory system. The influence of acute hypoxia on linear and nonlinear heart rate variability (HRV) has been studied, but the parameters in the process of hypoxia are still unclear. Although the changes of HRV in frequency domain are related to autonomic responses, how nonlinear dynamics change with the decrease of ambient atmospheric pressure is unknown either. Eight healthy male subjects were exposed to simulated altitude from sea level to 3600 m in 10 min. HRV parameters in frequency domain were analyzed by wavelet packet transform (Daubechies 4, 4 level) followed by Hilbert transform to assess the spectral power of modified low frequency (0.0625-0.1875 Hz, LFmod), modified high frequency (0.1875-0.4375 Hz, HFmod), and the LFmod/HFmod ratio in every 1 min. Nonlinear parameters were also quantified by sample entropy (SampEn) and short term fractal correlation exponent (alpha1) in the process. Hypoxia was associated with the depression of both LFmod and HFmod component. They were significantly lower than that at sea level at 3600 m and 2880 m respectively (both p < 0.05). The LFmod/HFmod ratio was acutely increased at 3600 m (p < 0.05). SampEn was significantly declined at 2880 m (p < 0.05). Although the value of alpha1 was close to 1, it changed not significantly in the whole process. These results indicated hypoxia gradually attenuated both spectral HRV parameters and SampEn. The balance of sympathovagal shifted towards sympathetic dominance at a certain altitude. Monitoring linear and nonlinear HRV parameters continuously in the process of hypoxia would be an effective way to evaluate the different regulatory mechanisms of autonomic nervous system.