The typology analysis of the spectral density curves of time series obtained from various parame ters of unit pulse oscillations was performed. The formal classification (typology) of spectral curves was per formed using correlations of peak values of the spectral density of VLF, LF, and HF slow waves. To include a spectral density curve into a particular type, the criteria were introduced that characterize the intensity of these oscillatory components. The experimental data was collected during examinations of children and ado lescents that were performed to detect the early stage of arterial hypertension (AH). The distribution of the subjects among the types of spectral density based on various parameters of pulse oscillation was evaluated. The results of the comparative analysis of spectral type distribution for various parameters are shown. This analysis has revealed the most informative spectral parameters for differential diagnosis of conditions of two types: type 1, essential hypertension, and type 2, various functional disorders with a normal blood pressure. It has been found that the parameters of the oscillatory components of dicrotic pulse wave time characteristics are the most informative for detecting hypertension in children.