In order to improve the evaluation precision of light environment for ship night navigation, so as to better control the light pollution at sea, and ensure the safety of ship’s night navigation, on the basis of considering the fuzziness and uncertainty of boundary in the classification of light environment quality grade, a light environment evaluation model for ship night navigation was proposed based on improved cloud matter-element structure by using the uncertain reasoning characteristics of cloud models and both advantages of quantitative and qualitative analysis of matter-element extension theory. The analytic hierarchy process (AHP), entropy weight and mean square variance (MSV) were used to determine the weights of light environment evaluation indexes, thus weakening the influence of subjective factors. The standard cloud matter-element model of light environment quality was constructed, and the degree of association between the evaluated matter-element and the standard cloud matter-element model was calculated. The light environment quality grade was determined by the information entropy theory. The confidence coefficient was defined by using theory of information entropy, which makes the reliability of evaluation results easy to calculate quantitatively. The validity and superiority of the model are verified by the practical example.