Low complexity distributed active noise control (ANC) systems are highly demanded for industrial applications. The ANC systems based on centralized multichannel filtered s least mean square (FsLMS) and Chebyshev-functional link artificial neural network (C-FLANN) FsLMS have shown promising results in cancellation of active noise. In distributed ANC scenario, there is a possibility in reducing the complexity by incorporating constraints on secondary paths. In this paper, we present improved distributed C-FLANN FsLMS algorithm for active control of nonlinear noise process. Based upon simulation study, it has been proven that the computational complexity can be significantly reduced by considering the error introduced at the target node due to the neighboring nodes in addition with self error of that node. Without significantly degrading the noise cancellation performance, the improved constrained distributed C-FLANN FsLMS scheme can achieve complexity improvement of 33.86% for multiplications and 22.58% for additions as compared with multichannel FsLMS considering 4 nodes. © 2016 IEEE.