Quality of Service (QoS) prediction and aggregation for composite services is one of the key issues in service computing. Existing solutions model service QoSs either as deterministic values or probabilistic distributions. However, these works overlooked an important aspect in QoS modeling, time. Most QoS metrics, such as response time, availability, are time-dependent. We believe time variation should be explicitly reflected in QoS modeling as well as aggregation. In this paper, we propose a dynamic web service QoS model to capture the time based QoS patterns, based on which QoS of composite services are aggregated.