Location information is crucial for many applications of sensor networks to fulfill their functions. A mobile sensor network is comprised of both mobile and stationary sensors. So far little work has been done to tackle the mobility of sensors in localization for sensor networks. In this paper, we propose the QoL-guided distributed refinements for anchor-free localization in wireless mobile sensor networks. Accuracy is the core concern for localization. We introduce the important concept of Quality of Localization (QoL) to indicate the accuracy of a computed location for a specific sensor node. Our approach is divided into two phases. In Phase one, we propose the algorithm QoL-guided spreading localization with refinements to compute locations for sensor nodes right after the deployment of the sensor network when the mobile sensors are required to stay static temporarily. In Phase two, the non-movement restriction is released and we propose the mobile location self-updating algorithm to update locations of mobile sensors regularly or on demand. Extensive simulations are conducted, which demonstrate that our approach is a promising technique for localization in wireless mobile sensor networks.