the continuous release of multi-point sources is one of the most common cases in the field of air pollution. In order to solve the problem of multi-point sources continuous release inversion, a source intensity and location estimation method based on improved Four-Dimensional Variation is proposed. Firstly, by constructing the diffusion equation of multi-point sources continuous release and the monitoring concentration matrix, the source term inversion model of multi-point sources continuous release is formed. Secondly, the joint transformation method and Taylor series upwind difference method are used to solve the convection-diffusion equation of multi-point sources continuous release, and the numerical solution format of the forward problem is formed through simulation analysis. Furthermore, based on the numerical solution scheme of the forward problem, a four-dimensional variation inversion algorithm for multi-point sources continuous release is constructed, and the flower pollination algorithm is used to improve the inversion accuracy and computational efficiency. Finally, the applicability, feasibility, advantages and disadvantages of the improved four-dimensional variation algorithm are verified by numerical simulation analysis. It is found that the improved four-dimensional variation algorithm can realize the source term inversion under three conditions: the location is known and the intensity is unknown, the intensity is known and the location is unknown, the intensity and the location are all unknown, which the accuracy and computational efficiency can basically meet the actual needs.