税收在国家治理中发挥着基础性、支撑性作用，实现智慧税务是政府在数字时代转型的必然要求，因而梳理智慧税务中的关键问题并探讨发展思路兼具理论研究与实践应用价值。本文分析了我国智慧税务领域的发展现状及面临挑战，提出了以“数据知识化、知识体系化、知识可推理”为核心的大数据知识工程解决方案，构建了由知识源层、知识提取层、知识图谱层、知识推理层、应用层组成的“五层”技术架构；结合大数据知识工程在智慧税务领域中的代表性应用案例，如知识驱动的税收优惠计算、可解释的税收风险识别、税收政策智能化决策支持、智慧问税，探讨了所提方案的局限性并论述了进一步的研究方向。从数据、技术、生态三方面出发，形成了规范涉税数据、健全国家数据共享/开放/保障体系，融合并更新信息学科成果，完善面向智慧税务的大数据知识工程应用系统，推动大数据知识工程技术的标准建设与人才培养等发展建议，以期为基于大数据知识工程的智慧税务高质量发展研究提供参考。

Taxation is vital for national governance, and the digital transformation of governments necessitates smart taxation. Therefore, analyzing the key issues and exploring the development ideas for smart taxation is of both theoretical and practical values. In this study, following an analysis of the development status and challenges facing China's intelligent taxation field, we proposed a big data knowledge engineering approach that emphasizes data knowledgeization, knowledge systematization, and knowledge reasonability, and developed a five-layer technical architecture that consists of knowledge sources, knowledge extraction, knowledge mapping, knowledge reasoning, and application layers. After elaborating the representative application scenarios including knowledge-driven tax preference calculation, interpretable tax risk identification, intelligent decision support for tax policies, and smart tax questioning, we investigated the limitations of the proposed approach and further discussed the directions for future research. Furthermore, we proposed the following development suggestions in terms of data, technology, and ecology: (1) standardizing tax-related information and improving the national data sharing, opening, and guarantee system; (2) integrating the achievements of various information disciplines and improving the application system of big data knowledge engineering for smart taxation; and (3) promoting talent training and the development of technical standards for big data knowledge engineering.