Abstract: A thresholding wavelet multiresolution wavefront reconstruction algorithm is proposed in view of low calculating efficiency and large stable state error of conventional wavefront reconstructors. The wavefront phase distortion is projected to wavelet basis, which makes the pseudo-inverse matrix become sparse matrix and reduces the calculation. Meanwhile, threshold is chosen to remove the high resolution wavelet-based coefficients which have little effect on calculation precision, improving operation efficiency. The algorithm is compared with least square, Zernike modal and iterative methods, and the simulation results show that this algorithm is superior in computing speed and convergence precision.