The march in patients affected by Parkinson's disease (PD) is characterized by pronounced corporal imbalance and instability of movements. This work studies new parameters that discrimine between Parkinson's and healthy people, using the analysis of the march signals, specifically the length between two consecutive steps. The parameters analyzed from the time interval vector of steps were: the variability coefficient and the useful energy bandwidth of the movement over -10dB of the estimation of the signal's spectral power through periodogram These parameters were evaluated in 15 PD patients and 15 healthy people (control group). Statistical T-student test for both parameters shows that the useful bandwidth (p=0,0233) was differences statistically significantives between both groups. Meanwhile for the variability coefficients was not differences statistically significantives between groups (p=0,346177667). The proposed procedure can be used as basis of research in early alterations march detection, and could be used to precocious Parkinson's diagnosis. However, this study does not discriminate among the diverse diseases that affect the march, it is recommended to carry out news researches increasing the number of signals and population.