The　maintenance　decision　needs　to　consider　various　aspects　such　as　cost,　safety,　system　condition　etc.　This　paper　presents　a　condition-based　maintenance　optimization　model　with　a　risk　acceptance　criterion　under　imperfect　inspection.　The　inspection　interval　can　adaptively　change　with　the　system　condition.　In　order　to　make　the　maintenance　decision　more　reasonable,　the　optimization　model,　which　minimizes　the　expected　long-term　cost　rate　with　safety　constraints,　also　considers　the　influence　of　imperfect　inspection.　The　method　is　illustrated　through　the　case　of　a　feeder　pipe　of　a　nuclear　power　plant.　The　results　show　that　the　safety　constraint　and　measurement　error　cannot　be　ignored.