Laser Doppler Flowmetry (LDF) is a powerful tool to explore, non-invasively, in vivo skin microcirculation. LDF produces a complex signal with oscillatory and fractal properties, sometimes difficult to evaluate. This study aims to characterize the oscillatory components of the LDF signal by mathematical modelling during a classic ‘oxygen challenge test’. A group of 35 healthy subjects (22.1±3.7 years old), both genders, were enrolled after giving their written informed consent. The measurement procedure included: baseline measurement, while breathing room atmosphere; provocation by breathing a 100% oxygen atmosphere; and recovery, while resuming normal atmosphere breathing. LDF signal (PF5010 system, Perimed, Sweden) was recorded on the inferior side of the second toe and then decomposed by: the wavelet transform, which rendered its components’ activities (cardiac, respiratory, myogenic, sympathetic, and endothelial); detrended fluctuation analysis (DFA) which allowed the characterization of the self-similarity properties through its alpha (α) exponent; and multiscale entropy analysis (MSE), quantifying the signal complexity over multiple time scales. All statistical comparisons were done with the Wilcoxon signed-rank test (p<0.05). The hyperoxia led to a significant perfusion reduction (p=0.001) followed by full recovery. The endothelial and sympathetic components contributed the most to the LDF signal. During provocation, a significant increase in the respiratory activity (p=0.006) was noted. Cardiac and myogenic activities also increased, while sympathetic and endothelial activities decreased, all of which showing no statistical significance. Wavelet components showed α>0.5, meaning positive self-correlated signals. On baseline, cardiac and respiratory activity components have shown α~1.00, suggesting that these phenomena have a 1/f noise-like behavior. However, myogenic (α~1.44), sympathetic (α~1.54) and endothelial (α~1.47) activity components reflected other characteristics closer to the Brownian noise (α=1.50). α values decreased for all, except for the endothelial component. The sole α value significant change was noted for the cardiac component (p=0.02), which also exhibited the highest entropy level, suggesting a more random-like signal. The components’ entropy levels changed, although non significantly, during provocation – cardiac and endothelial levels decreased, while respiratory, myogenic and sympathetic levels increased. These results suggest that the combined use of the wavelet transform, DFA and MSE contribute to characterize the complex LDF signal and thus helps to better understand the different mechanisms underlying peripheral perfusion regulation.