Infrastructure-as-a-Service (IaaS) cloud providers offer diverse pricing plans, namely on-demand, reservation, and spot pricing plan to meet distinct preferences of customers. This allows cloud users to purchase cloud instances based on their own demand and to generate less cost accordingly. In this paper, we propose a model of multi-instance combination decision for minimizing customer’s cost under cloud demands fluctuation with taking risk cost into consideration. We discuss the condition and reason of multi-instance optimal decision of cloud customers. And through case, we analysis the relationship between the optimal decision and fluctuation forms, fluctuation range, effective working time and risk sensitive factor of cloud users’ future demand respectively. Our extensive simulations driven by large-scale Google cluster-usage traces have shown that this model can minimize customer’s cloud customer’s cost which bring a series of economy effect on cloud computing investment of enterprises.