Abstract: Traditional sandbox simulation aids in understanding complex terrain and spatial relationships, facilitating observation and analysis. In recent years, sandbox simulation technology has primarily relied on planar similarity simulation, which suffers from issues such as poor accuracy and low automation levels. Consequently, 3D-printed simulation sandboxes have emerged, demonstrating advantages in three-dimensional simulation. While 3D similarity simulation can, to some extent, address the shortcomings of planar similarity simulation, existing 3D printers for simulation sandboxes exhibit inadequate pressure control, leading to problems like over-stacking and interruptions. Based on China's first sandbox 3D printer, improvements were made to the upper structure by adding one degree of freedom to achieve diversification of printed models. Furthermore, a pressure control system based on visual inspection was designed. This system utilizes a Basler camera for visual inspection and image processing to analyze printing quality and effectiveness during the process, while also monitoring the pressure within the material delivery tube. A fuzzy PID control-based pressure control system was designed for the 3D printing sandbox, resolving issues such as interruptions and over-stacking caused by pressure instability during printing. Experimental results indicate that the vision-based pressure control system for the 3D printing sandbox can maintain the tube pressure within the optimal printing range more effectively, increasing the qualification rate of maintaining pressure within the optimal range by 24.25% during the printing process. The design of this system enables accurate and continuous output for the 3D printing sandbox, meeting its performance requirements.