Orthogonal matching pursuit (OMP) algo rithm of compressive sensing should take sparsity as a priori condition,but the a ctual sparsity is not easy to be determined.In this paper,the method to determin e sparsity and twice orthogonal matching pursuit algorithm (TOMP) are proposed.F irstly,the entropy-weight method is introduced to determine the sparsity with c ombination multi-index fusion and the method of saturation point,and then the p roposed method applied to reconstruct the signal.Simulation results show that,compared to the similar algorithm,the signal processed by the proposed algorithm in creased 0.1second to reduce 12～22% reconstructon error;while compared with diff erent algorithms,signal noise ratio (SNR) of the si gnal after processing by the proposed method can be improved by 22dB,and root m ean square error (RMSE) can be reduced by 0.7,so i t demonstrates better de-noising effect.