High-fidelity deception jamming can seriously mislead Synthetic Aperture Radar (SAR) image interpretation and target detection, which is difficult to identify or eliminate through traditional anti-jamming methods. Based on the Range-Doppler Algorithm (RDA), an anti-jamming method for SAR by using joint waveform modulation and azimuth mismatched filtering is proposed in this paper. The signal model of SAR echo after pulse compression with deception jamming is derived from the radar detection and jamming characteristics. A multi-objective cost function is introduced to suppress the jamming level and keep the real target information, parameterized with the waveform initial phase and the azimuth mismatched filter coefficients, which is optimized using the second-order Taylor expansion approximation and the alternating direction multiplier method (ADMM). The performance of this proposed method is evaluated through the convergence analysis and anti-jamming experiments on the point target and the distributed target scenarios.