The invention belongs to the technical field of flow measurement, and particularly relates to a multiphase flow online real-time metering artificial intelligence model construction method. Oil-gas-water multiphase flow physical property parameters are various, and the flow process is complex, so that the development of a real-time flow measurement technology in an oil-gas production process is seriously restricted. The invention provides a multiphase flow online real-time metering artificial intelligence model construction method. The method comprises the steps of adopting Venturi measurementequipment, an electrical tomography sensor and a microwave sensor to cooperatively measure signals; by means of a deep encoding-decoding mode, under the condition that a large number of real flow sample tags do not exist, fully utilizing a large number of flow-tag-free measurement signals, and extracting the most basic constituent unit signals in the measurement signals; and constructing an artificial intelligence identification model of the flow of the measurement signal, the gas phase, the oil phase and the water phase according to the extracted minimum measurement signal forming unit. And meanwhile, the label-free measurement signal is fully utilized, so that higher measurement precision is realized.