The　problem　of　image　reconstruction　based　on　compressed　sensing　using　single　layer　wavelet　as　a　sparse　transformation　is　discussed　in　this　paper.　An　image　reconstruction　method　based　on　Smoothed　L0　(SL0)　algorithm　is　presented,　in　the　view　of　increasing　the　quality　of　image　reconstruction.　The　simulating　results　show　that,　comparing　with　primitive　orthogonal　matching　pursuit　(OMP)　algorithm,　this　reconstruction　method　can　significantly　improve　its　performances,　including　image　reconstruction　quality　and　robustness,　as　well　as　executing　speed.　©　2012　IEEE.