All experiments of pressure fluctuations were carried out in a bubble column with a moderately large column of 0.376 m ID. The recently developed technique of wavelet packet transform based on localized wavelet functions is applicable to analysis of the fluctuating signals. The time series of pressure fluctuation signals have been analyzed by means of wavelet packet transform components, decomposition through best basis algorithm and time-frequency representation. By resorting to this technique, the objects in bubbly flow regime have fine scales and frequencies than ones in churn-turbulent flow regime. Thus, this wavelet packet transform method enables us to obtain the frequency content of local complex flow behaviors in a bubble column.