In this work, the chaotic bubbling mechanism in a gas-liquid bubble column with a single nozzle was investigated. The signal for the analysis was the time series of pressure fluctuations measured from a pressure transducer probe placed in the bubble column close to the nozzle. In order to study the bubbling process, statistical analysis, qualitative and quantitative nonlinear analyses were carried out for the pressure fluctuations. Power spectra used as standard statistical measures provided preliminary evidence that bubbling in the middle values of gas flow rates may be chaotic in nature. Phase plots provided a qualitative means of analyzing the fine geometry structure of the attractor reconstructed from the bubbling time signal. Positive finite estimates of the Kolmogorov entropy provided a quantitative evidence of behavior consistent with chaos. Besides previous diagnostic tools, the local nonlinear short-term prediction was also used as a supplement method. It was found that the bubbling process exhibits a deterministic chaotic behavior in a certain range of the gas flow rate. When increasing the gas how rate, the sequence of periodic bubbling, primary and advanced chaotic bubbling, and jetting or random bubbling were successively observed. However, no clear period doubling sequence leading to chaotic behavior was observed. The sharp loss of the ability to predict the pressure signal successfully with the nonlinear prediction method provides the strongest evidence of the presence of the chaotic bubbling. The variations of the nonlinear invariants, such as the Kolmogorov entropy and the correlation dimension together with the plot of the correlation integral with the operation conditions, might be developed as potential and effective quantitative tools for flow regime identification of the bubbling process.