A remote sensing image JPEG2000 compression scheme with precise quality control was proposed. According to the perfect reconstruction of wavelet filter banks and the statistical independence of subband coefficient quantization error, a mathematical expression formula was established to describe the relation between the peak signal to noise ratio and the wavelet coefficient quantization error. Based on the traditional JPEG2000 rate distortion model, a distortion-based rate minimization model was constructed and optimization truncation coding stream to obtain a desired target image compression quality was given. An experiment on standard test images and satellite remote sensing image was performed and obtained results were compared with that of the conventional rate-based distortion minimization JPEG2000 encoding. Results indicate that the new method has the same coding structure and computing complexity, and it can accurately control the compression image quality with an average control precision less than 1%. In addition, this proposed method has higher global compression im-age peak signal to noise ratio at the same total coding rate for sequence remote sensing images.