Abstract: This study presents an integrated framework to minimize and quantify the deflection of thin steel plates during vacuum suction cup handling. The methodology combines parametric finite element simulation, space-filling design of experiments, and Gaussian process regression(GPR) surrogate modeling. Within the design space constrained by geometric and operational limits, 300 candidate suction cup layouts were generated via Latin hypercube sampling. Their structural performance was evaluated through finite element analysis to identify the optimal layout. Subsequently, the thickness sensitivity was analyzed by sampling plate thicknesses from 5 to 50 mm while maintaining the optimal layout. A GPR-based surrogate model was then constructed to map the thickness-deflection relationship, achieving a root mean square error of 0.22 mm on the validation set. Results demonstrate that the optimized layout reduces maximum deflection significantly, with deflection decaying approximately inversely with the square of thickness. The proposed framework efficiently accomplishes layout optimization and sensitivity analysis under a constrained computational budget, providing a practical tool for the design and safety evaluation of thin-plate lifting systems.