Current flight data processing research about data's inner attributes and interrelations between different data are lacking in depth. In this paper, theory and method of Takens are used to reconstruct the phase space of primal FDR flight data. The qualitative analytic methods, such as power spectrum figure and etc., are used to study the flight data. All the data curves are continuous and have wide peak characteristics. In the quantitative analysis, Wolf method on the basis of phase trajectory evolution is applied to estimating the largest Lyapunov exponents in different dimensions, and the results show that all the exponents are positive. So the chaotic characteristic of the flight data is verified. This conclusion is of novel reference value for the following-on research of this kind.