Abstract: To address the challenge of restricted manual access for close-range leak detection in future magnetic confinement fusion devices due to the intense radiation environment during the tritium operation phase, a remote leak detection strategy is proposed. This strategy integrates machine learning-based coarse localization using mass spectrometry data with fine localization via robotic helium spraying. However, given the scarcity of real-world leakage data and the absence of a validation environment for leak detection robots, this paper designs a remote vacuum leak detection test platform to facilitate the generation of simulated leak data and the verification of remote detection schemes. Constructed primarily from 304 stainless steel with a volume of 3.5\,\,\, \textm^3 , the platform integrates a two-stage pumping system—comprising a Roots-rotary vane unit for roughing and a magnetically levitated molecular pump for main evacuation—along with a 150℃ bake-out degassing system. This configuration achieves a working pressure of 5 \times 10^-5 \,\,\, \textPa , effectively simulating the high-vacuum environment required for fusion applications. The platform utilizes gas injection interfaces to allow the simultaneous introduction of multiple gas streams, simulating various leakage scenarios. Equipped with a Residual Gas Analyzer (RGA) and a full-range vacuum gauge, the system enables real-time monitoring of vacuum status and gas composition evolution, ensuring the accurate acquisition of simulation data. Static structural analysis of the main chamber was conducted using ANSYS software. The results indicate that the deformation, safety factor, equivalent stress, and stress intensity under vacuum negative pressure are all within safe limits, confirming that the structure meets mechanical stability requirements. Furthermore, theoretical calculations for the pumping and bake-out systems demonstrate a calculated ultimate pressure of 5.86 \times 10^-6 \,\,\, \textPa . The system is capable of reaching the working pressure within 20 hours, and the bake-out temperature can achieve the preset target. Finally, the fabrication and integration of the test platform were completed .