This research is dedicated for capital market analysis using wavelet method. Wavelet transform is widely used method in signal processing and image compression. Wavelets are ideally applicable for nonstationary signal analysis; consequently they are applicable also for financial time series analysis. Objects of experiment are 9 worldwide stock indexes: Dow Jones Industrial Average, AEX, CAC40, DAX30, IBEX, FTSE100, Nikkei225, SMI, STI. The Dow Jones Industrial Average index is considered for the whole available period: from 01/10/1928 to 03/10/2011, other mentioned indexes were analyzed during the following period: 05/07/1993 to 03/10/2011. Stock indexes are decomposed into trend and noise components using wavelet filtration. Both are analysed using direct continuous wavelet transforms. Interdependence between trend and noise components is shown. In this research an approach is worked out, which can identify instability in stock markets.