To address the problem of difficulty in distortion specttum demodulation of fiber Bragg grating (FBG) based sensing networks,we propose a wavelength demodulation technique based on estimation using a distribution algorithm (EDA). We construct a theoretical function of distortion spectrum based on the super Gaussian function and transform the wavelength detection problem of the distorted FBG sensing network into a function optimization problem. The proposed method is used to demodulate the distortion spectrum of a FBG sensing network through an experiment. The results denote that EDA can not only maintain an average detection accuracy within 1 pm even when the spectrum of FBG is distorted but also quantitatively estimate the distortion degree of FBG. When compared with the traditional peak detection methods, the proposed method can effectively identify the Bragg wavelength from a distortion spectrum. The proposed method provides a novel method to extend the service life and enhance the reliability of an FBG sensor network.