Abstract: To address severe interference from blended wavefields in high-efficiency simultaneous-source acquisition, this study optimizes multi-source encoding and proposes a dual-constrained encoding scheme in time and frequency domains. In the time domain, shot-to-shot dithering delays follow a uniform distribution and are subject to the constraints of non-repeatability, absolute threshold, and standard deviation threshold. In the frequency domain, based on the principle that blending noise maps to the high-wavenumber domain, the power spectral density trend is adopted as the criterion for evaluating encoding performance. Deblending experiments conducted on both synthetic and field datasets demonstrate that the proposed encoding scheme effectively reduces aliasing between signals and blending noise, thereby improving separation quality. Furthermore, this design principle exhibits strong robustness across different sparse inversion methods, confirming that its effectiveness is not limited to any specific separation algorithm.