自数据正式作为生产要素以来，其在生产领域的重要性日渐突出，数据资源对提高社会生产效率的作用日益凸显。与此同时，数据高价值、易受损的特性决定了政府需要通过制定有效的管理、监管和稽查制度来保障数据的安全和使用。税务稽查是政府保障税收公平公正、国家财政稳定的重要手段，税费数据是决定税务稽查质量的重要因素，在税务稽查中应用大数据治理技术已成为研究热点，尽管国内此类研究已经持续了20多年，但现行的税务稽查仍以较为传统的方式开展，仍未很好地应用大数据治理技术。本文深入分析了深圳市税务稽查工作现状以求探讨当前税务稽查应用大数据的困难及原因，基于“智慧稽查”的建设案例分析提出数据稽查的实施方式，并探究数据与业务整合后的稽查机制体系。在财务预算制度限制、信息系统林立、数据权属不清等背景下，提出数据稽查的具体实施路径，即在创新信息化项目管理办法下通过构建中台、集成现有系统、统一数据规范、开展数据确权与认证，应用新兴技术多维度校验的方式实施。本文通过比较分析数据稽查和税务稽查在目的、依据、对象和手段等方面的异同点，得到数据稽查的共性和特性。与传统的要素稽查相比，数据稽查要求更低的人力成本和更高的复杂运算。数据与业务整合后的稽查机制特征更为突出：整合机制下的实施主体、稽查惩罚方式与业务稽查的主体高度相关；整合机制下的数据稽查发生时间相对前移；相比于整合前的单一稽查方式，该机制的主要优势是业务稽查与数据稽查可以形成互补闭环，数据稽查的结果需要业务稽查等实操性行为进行验证，而业务稽查的成功与经验能够催化一批新的数据稽查，实现互相促进和优化，有效提高数据与业务稽查效果和质量。通过本研究将为税务稽查中的大数据应用难题提供解决思路，为其他领域数据稽查的开展与整合提供范本与经验。

Since data became an essential factor of production, its importance in the field of production has become increasingly prominent. The role of data resources in improving social productivity efficiency is becoming more and more evident. At the same time, The high-value and vulnerable characteristics of data determine that it is necessary for the government to ensure the security and use of data by formulating effective management, supervision and audit systems. Tax audit is an important means for the government to ensure tax fairness, justice, and national financial stability. Tax data are an important basis for determining the quality of tax audit. The application of big data governance technology in tax audit has become a research hotspot. Although such research has been ongoing for more than twenty years, the current tax audit is still carried out in a relatively traditional way, and big data governance technology cannot be well utilized.This article analyzes in-depth the case of tax audit projects and the construction of "Smart Tax Audit " in Shenzhen, aiming to explore the difficulties and reasons for the application of big data in current tax audit, then propose solving existing problems through data audit, and study the mechanism system of data audit and business integration. Under the conditions of financial budget system limitations, numerous information systems, and unclear data ownership, the implementation path of data audit is proposed, that is, to build a middle platform, integrate existing systems, unify data standards, carry out data rights confirmation and authentication, and implement multi-dimensional verification using emerging technologies under innovative information project management methods. By comparing and analyzing the similarities and differences between data audit and tax audit in terms of purpose, basis, objects and means, the commonalities and characteristics of data audit can be obtained.Compared with general element audit, data audit has lower labor costs and higher complexity of operation. The advantages of the audit mechanism for integrating data and business are more prominent, the subject of the mechanism and the punishment of audit are highly related to the business audit entities, the relative occurrence time of the data audit moves forward. Compared with the single audit mode before integration, the main advantage of this mechanism is that the business audit and data audit can form a complementary closed loop in terms of characteristics, and the results of data audit need to be verified by practical operational behavior such as business audit. The success and experience of business audit can catalyze a new batch of data audits, realizing mutual promotion and optimization, and improving the audit effect and quality. This study will provide solutions to the big data governance application challenges of tax audit and provide a model and experience for the development and integration of data audit in other fields.