There are various approaches for QoS-based web service composition. However, most of them are concerned about the algorithms of service compositions while ignoring the flexibility and expressiveness of users to set QoS constraints. Most of them assume that users could specify accurate QoS constraints easily. In reality, it is often not true especially for non-expert users. Users may not know the exact ranges or values of QoS requirements for their tasks. They may just want service composition solutions based on current QoS technical levels or want composition solutions in cost priority or quality priority. To deal with this non-clarity and variety of QoS requirements, in this paper, we propose a multi-strategic approach of service composition. This approach provides four strategies and aids to help users complete their QoS constraints and find optimization web service composition solutions.