In　this　paper,　a　method　for　predicting　remaining　useful　life　(RUL)　of　turbine　blade　under　water　droplet　erosion　(WDE)　based　on　image　recognition　and　machine　learning　is　presented.　Using　the　experimental　rig　for　testing　the　WDE　characteristics　of　materials,　the　morphology　pictures　of　specimen　surface　at　different　times　in　the　process　of　WDE　are　collected.　According　to　the　data　processing　method　of　ASTM-G73　and　the　cumulative　erosion-time　curves,　the　WDE　stages　of　materials　is　quantitatively　divided　and　the　WDE　life　coefficient　(ζ)　is　defined.　The　life　coefficient　(ζ)　could　be　used　to　calculate　the　RUL　of　turbine　blades.　One　convolutional　neural　network　model　and　three　machine　learning　models　are　adopted　to　train　and　predict　the　image　dataset.　Then　the　training　process　and　feature　maps　of　the　Resnet　model　are　studied　in　detail.　It　is　found　that　the　highest　prediction　accuracy　of　the　method　proposed　in　this　paper　can　be　0.949,　which　is　considered　acceptable　to　provide　reference　for　turbine　overhaul　period　and　blade　replacement　time.　Copyright　©　2020　ASME