Abstract: Aiming at the problems of such as the destruction of visual resources, model development and lack of characteristic features of waterfront greenway landscape, from the perspective of visual attraction, taking the Zhengzhou Dongfeng canal waterfront greenway as an example, the basic data of 10 indicators such as spatial scale, plant, entity, color, texture, water, dynamic scene, landscape diversity, landscape naturalness and landscape coordination were obtained by OpenCV and SBE methods respectively, then combined with the entropy method to calculate the weight,the factor correlation and stepwise regression analysis were made by SPSS 25.0 software. Finally, the visual attraction model was established and the greenway was evaluated. The results showed that the visual attraction quality of Dongfeng Canal's waterfront greenway was high, Wenbo East Road-Garden Road Section and North Third Ring Road-Xuzhai Road Section were the highest, and Dongming Road-Zhongzhou Avenue section was the lowest, and the difference was significant.

Key words: landscape space, waterfront greenway, visual attraction, entropy weighting method, Dongfeng canal