Pressure fluctuations in gas-solid fluidized beds have long been recognized as an important index for the quality of fluidization. Due to their highly random nature, many researchers have resorted to stochastic signal-processing methods to extract useful information. However, methods such as power-spectral density are based on a priori hypothesis of the stationarity of the signals, whose validity has not been investigated. In this article, using the Wigner distribution (WD), the signal’s strong nonstationary feature is demonstrated and signal properties are addressed. Local frequency information is extracted from the defined parameter-the local peak weighted average (LPWA), which can be regarded as a generalized concept of the major frequency of power-spectral density in the nonstationary case. Experimental results indicate that LPWA is stochastically much more reproducible than the major frequency of power-spectral density. The change in LPWA under different operating conditions may reflect the corresponding change in the bubble phase. This information is useful in characterizing the quality of fluidization.