Cloud computing is a promising solution to provide the resource scalability dynamically. In order to support large scale workflow applications, we present Nuts-LSWAP which is implementation for Cloud workflow. Then, a novel Min-cost with delay scheduling algorithm is presented in this paper. We also focuses on the global scheduling including genetic evolution method and other scheduling methods (sequence and greedy) to evaluate and decrease the execution cost. Finally, three primary experiments divided into two parts. One parts of experiment demonstrate the global mapping algorithm effectively, and the second parts compare execution of a large scale workflow instances with or without delay scheduling. It is primarily proved the Nuts-LSWAP is efficient platform for building Cloud workflow environment.