In order to make the methods and strategies of artificial intelligence with data-driven and massive historical operation data combine directly with each other, and then meet the intelligent demand of multi-objective combustion optimization for coal-fired power stations, an intelligent historical database (IHDB), which is a combination of artificial intelligence and historical database, is proposed. The structures of file, index and module, which compose the kernel of IHDB, are presented as well. The kernel design doesn’t only ensure processing data efficiently, but also incorporates data analysis, optimization decisions and other intelligent elements, so that a strong platform can be developed for solving the problem of multi-objective combustion optimization. The performance tests about storing and querying data of IHDB have obtained satisfactory results. Developing IHDB is beneficial to improve the intelligent level of coal-fired power plants and is helpful to promote the development of real-time database system involving IHDB.