Based on Electrical capacitance tomography (ECT), a softsensing model used to identify flow regimes of gasliquid twophase flow is proposed.Using neural network technique, a mathematic relation between the second variable (the character parameters extracted from ECT sensor outputs) and primary variable (the flow regime of gasliquid twophase flow) of the softsensing model is established. Then the flow regime of gasliquid flow can be identified online. The simulation result shows that the proposed method has good identification precision and fast identification speed, which means it is an effective tool in twophase flow pattern online identification.