Abstract: For solving the cost optimization of the scientific workflow scheduling in cloud environment, a scientific workflow scheduling algorithm CWS_PSO based on improved particle swarm optimization (PSO) model is presented in this paper. Our algorithm formalizes the scientific workflow scheduling problem as the scheduling cost optimization problem with the deadline constraint. The encoding and transfer model are constructed between the particle's position and tasks scheduling scheme by PSO model, and searching for the cost optimization scheme satisfying the deadline constraint is processed by the particle's iteration evoluation. Through the simulation experimental of scientific workflows model, the performance comparisions are conducted in the same kind of algorithms. Experimental results show that our algorithm could efficiently achieve the trade-off between the workflow execution makespan and the execution cost, and achieve the execution cost minimization with the deadline constraint satisfaction.