Capacity　is　an　important　property　of　digital　watermarking.　Research　of　image　watermarking　capacity　is　to　study　how　much　information　can　be　hidden　in　an　image.　Recently,　wavelet　transform　has　been　applied　widely　in　watermarking　research　as　its　excellent　multi-resolution　analysis　property.　This　paper　realizes　a　content　adaptive　watermarking　algorithm　using　Watson　quantization　matrix　and　Noise　Visibility　Function　(NVF),　and　discusses　the　capacity　problem　of　watermarking　in　wavelet　domain.