Data stream, a new kind of data model, is proposed recently to solve continous queries problem. with data streams has attracted more and more attentions, lots of researchers have proposed all kinds of algorithm for different applications. however, most of the algorithms are designed for single processor. when many streams arrive at very high rates, the system will potentially lost many result when the CPU is saturating. obviously the single processor will become a nottleneck, and a multi-processor provides a good choice. this paper proposes a research on the optimization of multi-join queries in multi processors. in that scenario, one of the key problems is how to arrage the streams's probe sequences. two algorithm are presented in this paper. one is heuristic-adjusting algorithm with the minimum spanning tree as the heuristic rules, which is suitable for the scenario that the input stream's rate are changing slowly. the other is a static-solving algorithm using the idea of greddy algorithm, which is suitable for the scenario that the input stream's rate almost unchanged. the experimental result show that the two algorithms can achieve the goal that improving the throughput to a certain extent.