To address the issue of the transmission characteristic curve of electro-optic intensity modulators being prone to drift due to environmental influences，this paper proposes a bias control method based on an improved particle swarm optimization algorithm. Based on the structural characteristics of the Mach-Zender electro-optic intensity modulator，the transmission characteristic curve of the modulator is derived. The effects of different bias voltages and RF amplitudes on modulation performance are simulated and analyzed. The optimal linear working point of the Mach-Zender electro-optic intensity modulator is determined to be ±Quad point. Then the principle of traditional particle swarm optimization is introduced. In order to optimize the performance of particle swarm optimization in bias automatic control problems，improvements are made to the particle swarm algorithm by dynamically adjusting inertia weights and particle swarm individual distribution，constructing specific fitness functions，and optimizing multi-objective parameters. Finally，the performance of the improved particle swarm optimization algorithm in bias automatic control problems is simulated and analyzed. It can accurately find the optimal linear working point and has a fast convergence speed. At the same time，the algorithm is transplanted to a self-developed bias control board. Comparative experiments are designed to prove that the improved particle swarm optimization algorithm can effectively track the drift of the bias curve. The amplitude change of the modulated output signal waveform does not exceed 2%，which meets the requirements of long-term use.