Future combats being extremely complex, and challenge to decision\|makers taken by significant cognitive overload due to the massive input of visual information, an evaluation method with key visual information characteristics, i.e. hue, saturation, value and image gradient as independent variables is proposed by the relevant theories of human visual perception and visual saliency mechanism. A quantitative model of visual search efficiency is constructed via a target search task over a complicated image map, and the visual search response time is used as a dependent variable. Three landforms (city, ocean, and plateau) consisting of 441 image maps with different contrast and brightness are tested in the psychology experiment, and the model result show that the R2 is equal to 0.232, p< 001. And the results have practical guiding evaluation and optimization to complex interfaces under different scenarios.