The nonlinear dynamics of in-line bubbles rising with coalescence in non-Newtonian Carboxymethylcellulose sodium (CMC) fluids was studied through the techniques such as the multiresolution signal decomposition and the chaotic time series analysis. The temporary signals of bubble passages collected by an optical sensing device at different heights were investigated by a 12-level wavelet decomposition and the scalewise characteristics of bubble motion were extracted and analyzed. The chaotic time series analysis distinguished the periodicity or the deterministic chaos of bubble motion successsfully. The calculation of Kolmogorov entropy proves that in the ranges of experimental heights and gas flowrates, the bubble rising dynamics becomes more chaotic with the increase of height, and reaches the maximum chaotic extent in a certain height, while with the further increase of height, the chaotic extent decreases slowly. With the increase of gas flowrate, at the lower height, the bubble rising dynamics changes from periodicity to deterministic chaos, and at the higher heights it reaches the maximum chaotic extent in a certain gas flowrate; however, for both cases, it has little change in the higher gas flowrates. Moreover, with the increase of CMC concentration, the bubble rising dynamics becomes less chaotic when the height is beyond a certain value.