This paper presents a wavelet analysis-based multi-resolution cone-beam volume CT breast imaging technique that is adaptive for high-resolution and ultra-high resolution reconstructions. Wavelet analysis-based de-noising techniques are employed to improve image quality and further reduce the required absorbed dose. The following steps can summarize this technique. First, in the high-resolution mode, the high spatial resolution projections are rebinned into lower resolution projections through a wavelet decomposition/synthesis procedure while in the ultra-high-resolution mode the original spatial resolution of the projection data is kept. Second, a wavelet analysis-based de-noising technique is applied upon the projection data with quantum fluctuations to suppress the noise level in the reconstructed images. Third, a de-noising method through an adaptation of a wavelet shrinkage approach for noise reduction is utilized in the reconstructed data to improve the image quality in terms of the signal-to-noise ratio and dose efficiency. The computer simulations show that the wavelet analysis-based multi-resolution rebinning approach provides the flexibility to adjust spatial the reconstruction resolution and noise level for various imaging tasks. Also, the wavelet analysis-based de-noising technique efficiently suppresses the quantum mottle induced noise, and contributes to a better low contrast object reconstruction in terms of the signal-to-noise ratio (SNR) improvement. In addition, the reconstruction of a high contrast object, for example, a tiny calcification grain, is obtained with less density spread. The noise level in the reconstructed image is reduced, which means the necessary dose level can be further reduced while the image quality is not compromised.