Symbolic time series analysis is a new tool for analysis of experimental data, it has been successfully applied in many fields. Results indicate that the symbolization can increase the efficiency of finding and quantifying information from dynamic systems, reduce sensitivity to measurement noise. Three statistical quantities were used to characterize symbolic time series, the method was validated by Henon equation. We also applied the methods to analyze the experimental data of oil/water two-phase flow in vertical upwards pipes and proved that the calculated statistical quantities from symbolic time series are sensitive to the transitional flow pattern variations of oil/water two phase flow.