A high robust approach for the additive spread spectrum audio watermarking scheme based on orthogonal decomposition of the host signal has been proposed. By analyzing traditional spread spectrum(SS), the section of the audio signal which is parallel to the direction of watermark signal acts as an interference source in decoder. The watermark is only embedded in the section which is perpendicular to watermark signal in the proposed scheme. In this way, the interference of host signal can be removed completely in decoder. Whether the watermarks have been embedded in the audio or not is researched and an optimal threshold in detector is given to achieve the minimum false alarm probability. The simulation results show that the proposed method can achieve high robustness against various attacks. A highly robust approach for the additive spread spectrum audio watermarking scheme based on orthogonal decomposition of the host signal has been proposed. By pars traditional spread spectrum (SS), the section of the audio signal which is parallel to the direction of watermark signal acts as an interference source in decoder. The watermark is only embedded in the section. which mark is only embedded in the section. or not is researched and an optimal threshold in detector is given to achieve the minimum false alarm probability. The simulation results show that the proposed method can achieve high robustness against various attacks.