In view of lower recognition rates of traditional methods in the flow pattern identification of Electrical Capacitance Tomography (ECT), a identification method for the ECT based on Compressed Sensing (CS) was put forward. Firstly, measurement capacitance vectors obtained by an ECT system were normalized and represented as a sparse linear combination of training sample set in an over complete dictionary. Then, the random Gaussian matrix was taken as the measurement matrix to sample from the test and standard samples respectively, and the signal reconstruction algorithm based on the CS was used to solve the optimization problem of L0 norm for the sparse representation of each sample on the training sample set. The linear correlation coefficient between the sparse solutions of samples to be tested and the standard samples are calculated to determine the classification of flow pattern. The simulation experiment results of typical flow patterns indicate that the flow pattern identification rates under absence of noise, and signal to Noise Ratios(SNRs) of 40 db and 20 db are 100%, 99.25% and 98.12% respectively. It concluds that the flow pattern identification method proposed has high efficiency and accuracy and a good noise immunity, which also provides a new method for the flow pattern identification of ECT.