In　the　whole　process　of　compressed　sensing　algorithms,　the　classical　sparse　dictionary　construction　usually　consumes　a　large　amount　of　computing　resource　and　shows　low　construction　efficiency.　As　a　result,　a　novel　compressed　sensing　diagnosis　method　based　on　a　single　mode　Laplace　sparse　dictionary　is　proposed.　A　characteristic　segment　containing　multiple　impacts　is　intercepted　from　the　long　signal,　and　a　sliding　window　is　applied　on　the　spectrum　of　the　segment　to　calculate　the　frequency　root-mean-square,　so　that　the　resonance　frequency　band　and　its　center　can　be　identified.　Therefore,　the　frequency　parameter　of　the　Laplace　wavelet　can　be　determined.　Furthermore,　a　sliding　time　window　is　applied　on　above　segment　and　the　shifting　kurtosis　index　curve　can　be　obtained　to　extract　the　single　impact　segment.　The　relative　filtering　method　is　then　utilized　to　determine　the　damping　parameter　of　Laplace　wavelet.　Zero　padding　interpolation　is　performed　to　establish　the　impact　atom　whose　length　is　equivalent　with　the　long　signal.　At　last,　the　cyclic　shift　strategy　is　applied　to　expand　the　single　atom　to　a　sparse　dictionary　matrix.　The　novel　dictionary　is　then　embedded　in　a　compressed　sensing　procedure　that　combines　a　Gaussian　measurement　matrix　and　orthogonal　matching　pursuit　algorithm.　As　a　result,　the　novel　compressed　sensing　diagnosis　method　is　founded　and　validated　by　simulation　and　engineering　data.　The　results　have　shown　that　the　single　mode　Laplace　wavelet　dictionary　can　improve　the　efficiency　of　dictionary　construction　and　reduce　the　storage　space　occupation.　Most　importantly,　the　compressed　sensing　procedure　integrating　the　single　mode　sparse　dictionary　can　obtain　better　compression　and　reconstruction　effects　than　traditional　methods,　which　can　ensure　the　accurate　fault　feature　identification　in　noisy　environments.　©　2024　Nanjing　University　of　Aeronautics　an　Astronautics.　All　rights　reserved.