Abstract: ［Purpose/Significance］As a novel feature of the web search engine to provide entity retrieval service and visualize the search results, the evaluation of knowledge card of search engines is beneficial for the improvement of services provided by search engines. ［Design/Methodology］With the use of the stratified random sampling method, 1200 entities were extracted from Wikipedia. By taking Google and Bing as examples, the evaluation test was conducted on the knowledge cards of search engines. 
［Findings/Conclusion］The total response rate of knowledge cards is less than 35%, and the rate of Bing is higher than that of Google. Generally speaking, users’ satisfaction is relatively high. Specifically, Google outperforms Bing in the indicators of relevance, while Bing has a better performance in the metric of comprehensiveness. In terms of content and structure of knowledge cards, the properties of Bing’s knowledge cards seem to be more abundant. Besides, only Bing provides links of knowledge resource, but the source seems relatively simple.［Originality/Value］This paper has proposed an evaluation method and corresponding metrics for the knowledge cards of Web search engines.

Key words: Search engine, Search engine evaluation, Knowledge card, Knowledge graph, Entity retrieval, Information retrieval visualization