The invention provides a memory management method used for parallel processing of streaming data. A processing line framework used for parallel processing of streaming data comprises multiple groups of producers and consumers in a peer-to-peer relationship in a one-to-one correspondence mode. Each producer and consumer in a peer-to-peer relationship is provided with a private local buffer pool. The memory management method used for parallel processing of streaming data comprises following steps of: 1), each producer applies memory to a private local buffer pool thereof during parallel processing of streaming data and triggers exchange operation in order to obtain a memory block when the local buffer pool of each producer is empty, wherein exchange operation is performed by exchanging a local buffer pool of each producer with a local buffer pool of a corresponding consumer; and 2), each consumer releases memory of a local buffer pool of the consumer during parallel processing of streaming data.The memory management method used for parallel processing of streaming data is capable of effectively decreasing cost of memory operation and increasing utilization rate of memory.