Guided wave (GW) is widely recognized as one of the most effective tools for structural health monitoring (SHM) due to their long-range propagation and high sensitivity to damage. However, the practical application of GW based SHM faces major challenges. Environmental and operational conditions, such as temperature and load, can cause variations in the amplitude and phase of GW signals. These variations may couple with damage induced changes, reducing the accuracy and reliability of conventional damage diagnosis methods. To overcome this challenge, this study proposed a GW feature signal extraction based damage imaging method under environmental and operational conditions. Firstly, multiple GW features are extracted from each path, which collectively enhance the characterization of structural damage from multiple perspectives. Probabilistic modeling is then employed to suppress the effects of environmental and operational conditions and constructed probabilistic GW feature signals, which are only sensitive to structural damage. By fusing the damage information of these signals through conventional damage imagining method, a reliable and accurate damage imaging can be achieved under environmental and operational conditions. To validate the proposed method, experiments were conducted on a typical aircraft structure. The results demonstrate that, compare with the conventional method, the proposed approach improves the accuracy and reliability of damage imaging across a wide temperature range from -50℃ to 80℃.