With its pronounced photon-fluctuation characteristics, super-thermal light provides a fundamental physical basis for overcoming the long-standing bottlenecks of sampling efficiency and image quality in conventional optical imaging. Building on previous demonstrations of super-thermal ghost imaging, this work integrates a super-thermal light source generated by secondary speckle modulation with compressive-sensing ghost imaging reconstruction into a unified imaging framework. Through two independent speckle modulations, a super-thermal light field with a g(2)(0) value enhanced from the thermal level of approximately 2 to nearly 4 is successfully generated, and the derived expression of g(2)(0) depends on the effective mode-number parameter and coherence length of the optical field. Under the sampling conditions studied here, this integrated approach improves the target–background separation and the quality of compressive-sensing reconstruction, demonstrating the application potential of engineered super-thermal statistics in advanced optical imaging.