Abstract: We decompose the hydrological sequence made of different frequencies into the low and high frequencies in the multi-resolution analysis according to the characteristic of hydrological sequence frequencies and then restore the deoxidize trend components according to the reconstruct principle of wavelet coefficients, in order to judge the hydrological sequence trend. The result of a study based on the gauged yearly flow of the Yellow River, shows that the mode is feasible.

Key words: evolution trend, time sequence, wavelet analysis, Yellow River