In , we presented a synthetic aperture radar (SAR) inversion method in which the backscatter measurements are collected from known but arbitrary flight path and non-flat topography corrupted with clutter and noise. The objective of the current work is to design transmit waveforms and receive filters jointly to perform clutter suppression by adaptive transmit. Our approach starts with the development of a physics based forward model for SAR. Next, we formulate design of clutter rejecting waveforms as a variational problem using the minimum mean square error criteria. The resulting waveforms lead to an adaptive linear least squares transmit scheme in which waveforms vary at every point on the flight path based on the clutter and target spectral density functions.