Aiming at improving the performance of obtaining texture regions by directional filters and achieving better steganography security, a novel adaptive steganographic algorithm is proposed based on WOW algorithm and Gabor wavelet. Firstly, a directional filter bank is established by Gabor wavelet and then the residual weights are determined by this bank from 8 directions. Then the cost function is defined based on H?lder norm, and is updated by convolution with an average filter. Finnally, secret messages are embedded by syndrome trellis codes according to the cost function. Experimental results illustrate that the presented steganographic algorithm achieves a better performance on resisting the Spatial Rich Mode steganalysis than that of the same kind of steganographic algorithms under the same secret message payload.