Imaging results under outdoor conditions are easily affected by foggy weather which may produce blur, low contrast, intensity shift and other visual degradation effects. This paper proposed an accessing model of fog density based on discrete wavelet transform features of multi-channel mean subtracted contrast normalized coefficients. Besides, corner feature intensity is utilized to extract high-level features of foggy images. The accessing model used support vector regression to predict fog density with the help of a fog density accessing dataset collected from human eyes. The experimental results show that the subjective accessing method proposed in this paper has a high correlation with human eyes accessing results on the testing dataset, and is better than the general referenceless image spatial quality evaluator model. The accessing model can reflect the image quality as human eyes in foggy scenes.