Owing to the problem that objective evaluation of fusion image quality is inconsistent with human visual effects in vehicle anti-halation, a new quality evaluation method is proposed in this paper for visible and infrared fusion images. The method can automatically determine halo critical gray values to divide fusion images into halo and no-halo regions by designing an adaptive iterative threshold algorithm. For the halo region, the halo limination effect is evaluated by designing a halo limination index. For the no-halo region, enhancement effects of color and detail information are evaluated from various aspects. Appropriate indexes are selected to constitute a complete image quality evaluation system. Fusion images of four different algorithms are evaluated to verify their rationality. Experimental analysis show that subjective and objective evaluation results of this method are consistent, rendering it suitable for evaluating anti-halation image quality and the algorithm of different visible and infrared fusion.