To improve the efficiency of MapReduce in heterogeneous cloud computing environment, the paper presents an adaptive scheduler based on workload type in heterogeneous environment (ASBT-HE). It feeds back results to the JobTracker to assign subsequent tasks to corresponding TaskTrackers. But in order to meet user’s QoS requirements and better balance the loads of system, ASBT-HE is optimized a better ASBT-HE (BASBT-HE).On the basis of considering the QoS constraints, BASBT-HE can adjust the length of queues in ASBT-HE automatically according to the actual loads. The ruslts of simulation experiments show that the optimized MapReduce has high efficiency in heterogeneous cloud computing environment. And it considers the user’s QoS requirements and simplifies the system configuration.