The　classical　vibration-based　structural　health　monitoring　evaluates　damage　via　the　analysis　of　single　type　measurement.　The　relationships　between　different　types　of　measurements,　however,　are　usually　neglected.　To　address　this　problem,　vibration　data　is　to　be　analyzed　in　state　space　in　this　paper　based　on　the　the　Poincare　map　method.　The　Poincare　map　method　is　an　effective　approach　in　damage　detection.　This　method,　however,　requires　the　dependent　measurements　for　displacement,　velocity　and　acceleration,　respectively.　Based　on　the　Fourier　transform　properties,　a　robust　Poincare　map　method　for　damage　detection　is　proposed　based　on　the　single　type　of　measurement.　Numerical　and　experimental　verifications　are　employed　to　verify　the　effectiveness　of　the　developed　method.　Compared　with　the　classical　numerical　differential　method,　the　present　method　is　more　accurate　and　robust　to　noise.