The invention discloses a streaming data real-time processing method and system based on a multi-level cache structure, and belongs to the technical field of big data stream processing. The method of the invention uses on-line data processing and batch processing for streaming data, so as to satisfy the requirements of low latency, high throughput and reliability of the processing. The method comprises the following steps: storing an on-line processing result in an on-line cache, and designing an on-line cache structure to automatically and efficiently delete stale data; storing data of a recent access offline persistent database in a batch cache, and adopting an LRU based strategy to efficiently eliminate the stale data; and merging a query result efficiently so as to return the query result to users quickly. The invention further realizes the streaming data real-time processing system based on the multi-level cache structure. The streaming data real-time processing method and the system based on the multi-level cache structure provided by the invention can solve the problems in the existing method that the storage and query efficiency is not high, the reliability is low and the memory occupancy rate is higher.