Aiming at the wavelength drift problem caused by the influence of ambient temperature on fiber bragg grating (FBG) strain sensor, this paper proposes the online prediction algorithm that combines particle swarm optimization (PSO) with sliding window extreme learning machine (SWELM) to compensate for temperature effect. The PSO algorithm is used to optimize the sliding window and the number of neurons in the hidden layer of the SWELM network, which improves the prediction accuracy of the model, and the minimum root mean square error of the model prediction can reach 0.06pm. PSO-SWELM realizes online update and wavelength drift prediction of strain sensor data, and differential calculation of real-time measurement data and prediction data completes temperature compensation. PSO-SWELM was compared with SWELM, and the results showed that the accuracy of the proposed algorithm was improved by an average of 11.04%, and has good temperature compensation effect