The image denoising variational method possesses higher numerical precision but lower computational efficiency, especially in processing the remote sensing images which is usually larger. To this problem, an adaptive 2-D wavelet interpolation operator was constructed based on quasi-Shannon wavelet function, and then adaptive wavelet precision integration method (PIM) on solving 2-D partial differential equation was proposed by combining with PIM. This proposed method combined the multi-scale property of the wavelet transformation with the higher precision of PIM, which could improve the computational efficiency effectively so that the lager remote sensing image could be processed.