Building a clean, low-carbon, safe, efficient and sustainable energy system has been listed as one of the national energy development strategies in China. The integrated energy systems (IES) integrates the processes of multiple energy generation, transmission, conversion, storage and distribution. The collaborative management and optimization for multi-energy resources is the key technology to raise energy efficiency, reduce costs and protect environments under a certain configuration of technologies and equipment, which provides the basis for forming a low-carbon sustainable energy operation mode, especially for the industrial parks. With the development of the big data and machine learning technologies, a series of data-driven methods have occurred in the field of the research on IES, covering the modeling, assessment and operational optimization of the IES, etc. The studies and issues of these aforementioned aspects are reviewed in detail, and the ongoing scientific problems and technical challenges that need to be further investigated are also presented. In addition, the future research directions of the data-driven methods for the IES operational optimization are discussed.