Abstract: Flow regime recognition plays an important role in petroleum multiphase flow research. However, the traditional flow regime identifying methods have many shortcomings. In this paper, the flow regime measurement techniques and the statistical analysis method such as probability density function and power spectral density are discussed. Additionally, the application of fractal, wavelet analysis and neural network are also presented. The methods applied in flow regime identification have been recommended and the emphasis during flow regime study has also been put forward.