To increase the redundancy of measurement data and improve the observability of linear harmonic state estimation, mixed measurements acquired from PMU and SCADA are used to build the mathematical model of non-linear harmonic state estimation. Then, it is rewritten as a sensitivity model, transformed into an optimization problem, and solved by particle swarm optimization algorithm. Example analysis shows the sensitivity model of non-linear harmonic state estimation is efficient. PSO algorithm can be used to solve this optimization problem. Mixed measurements are helpful to improve the accuracy of harmonic state estimation.