In order to achieve target type recognition when multiple targets are working simultaneously in fiber optic sensing network, a multi-target classification and recognition algorithm based on fiber optic sensing networks is designed. This algorithm constructs a solution model based on the wavelength response of test nodes, using signal amplitude, duration, and frequency as characteristic parameters of the target signal. Signal acquisition is carried out through a fiber optic sensing network networked with fiber Bragg grating (FBG), and the collected signals are subjected to feature extraction and analysis. The experimental results indicate that wavelength response tests are conducted on four typical vibration sources within a testing area of 20 m×30 m. The four targets have different signal characteristics: target 1 has a wavelength amplitude mean of 1 250 pm and a periodic feature of approximately 120 ms. The average wavelength amplitude of targets 2 and 3 is between 150~350 pm. The average wavelength amplitude of target 4 exceeds 3 200 pm, with a duration of approximately 15 seconds.