Abstract: Visual attention mechanism is introduced into the search process of special target(source of gas leakage)in order to improve the efficiency of visual information processing.On this basis,a novel search method of gas leakage source is proposed by using task-based visual attention mechanism(TBVAM).The method is divided into three steps.First of all, the combined weight coefficients of contrast maps that can efficiently highlight the target object in the scene are determined. Secondly,when information of the real scene's image is captured,a few suspected target areas are obtained by combining contrast maps of different features and scales with the trained weight coefficients,and the positions of the suspected targets are calculated with laser range information.Lastly,the robot approaches the suspected target areas and the gas leakage source is identified when the odor concentration detected by olfactory sensor is bigger than a given threshold.Experimental results show that the proposed algorithm can significantly improve the search efficiency of the gas leakage source.