Abstract: When using fiber Bragg grating (FBG) to measure traverse force, the reflective spectrum of FBG is severely influenced by the system errors, instruments wear and some other parameters hardly measured. It is very difficult to extract the characteristic information of transverse force. Therefore, an analysis method of FBG reflective spectrum under transverse force based on particle swarm optimization algorithm (PSO) was proposed, and a FBG transverse force-PSO model was established. The objective function was established for experimentalizing and optimizing reflective spectrum of FBG under transverse force, and then the characteristic information in reflective spectrum of FBG was extracted. Experimental data showed that PSO used in extracting the characteristic information is not only efficaciously and easily, but also has some advantages, such as high accuracy, stability and fast convergence rate, and it is useful in high-precision measurement of FBG sensor.