With the acceleration of smart grid construction, the requirements for extracting and analyzing the parameter information of power system to ensure the safe and stable operation of power system are getting higher. At present, a large number of overhead distribution line current in distribution system is difficult to achieve effective non-contact detection. To solve this problem, a non-contact three-phase overhead distribution line current detection method based on magnetic field inversion is proposed. Firstly, the inversion optimization model of line current is established according to the inverse magnetic field problem. Secondly, the non-dominated sorting genetic algorithm （NSGA-Ⅲ） with reference point mechanism is used to solve the model. The optimal solution that meets the requirements of decision makers is selected by combining the technique for order preference by similarity to ideal solution （TOPSIS） method, and the fitting curve is drawn and the characteristic parameters are extracted. Finally, the simulation analysis and experiment are used to verify the effectiveness and accuracy of the proposed inversion method.