In order to further research into the EEG information of rats epileptic model, we applied different nonlinear dynamic methods. After having analyzed the EEG signal of rat falling sickness by means of approximate entropy and correlation dimension, we adopted a the new method, unstable periodic orbits, which was used to analyze complex activity of neurons system to look for the change regularity of change in the EEG signal in the whole course of rat's falling sickness. We found period 1 orbits and period 2 orbits to be statistically significant in the data of ictal time of epilepsy. At the same time, we found period 1 orbits to be statistically significant in the data of preictal time of epilepsy.