Due　to　the　large　number　of　uncertain　factors　in　hybridization,　image　capture　and　processing　of　the　microarray,　multiple　probes　were　generally　arranged　to　improve　the　reliability　of　the　measurement.　However,　the　small　area　limited　the　number　of　probes　that　were　allowed　to　be　added　on,　so　a　composite　probe　would　be　the　better　choice.　A　composite　probe　contained　the　linear　combination　of　a　variety　of　gene　fragments.　It　was　used　so　that　the　microarray　could　easily　realize　the　repeated　gene　fragments　within　a　limited　region.　The　number　of　composite　probes　would　rapidly　dwindle　when　it　compared　to　a　traditional　microarray.　At　the　same　time,　since　the　sparse　characteristics　of　biological　gene　mutation,　the　compressed　sensing　idea　is　adopted　to　recovery　the　gene　variation　in　the　composite　probes.　The　96　fragments　can　be　used　with　the　48　x　96　sparse　random　matrix　to　construct　the　48　composite　probes　when　the　sparsest　level　K　is　no　more　than　12.　Simulation　results　show　that　compressed　sensing　can　accurately　recover　the　gene　mutation　by　using　the　Orthogonal　Matching　Pursuit　(OMP)　algorithm.