This paper presents the development of a neuro-active control mechanism for broadband noise cancellation and vibration suppression. A neural network architecture based on an orthogonal forward regression algorithm is developed. An active noise control (ANC) structure is considered for optimum cancellation of broadband noise in a three-dimensional propagation medium. An on-line adaptation and training mechanism allowing the neural network architecture to characterise the ideal (optima)) controller within the ANC system is developed. The neuro-adaptive active control algorithm thus developed is implemented within a simulation environment characterising the ANC structure and simulated results verifying the performance of the algorithm are presented and discussed.