<p>With the advancement of phased array radars toward high-performance, high-frequency, and highly integrated configurations, accurate prediction of antenna assembly precision has become a critical challenge. Conventional prediction methods, constrained by prohibitive computational costs, demonstrate limited applicability to engineering precision assemblies. To overcome this limitation, the present study proposes a novel methodology for assembly accuracy prediction based on surface contact potential energy optimization. The methodology begins with constructing a three-dimensional mating surface model using the Weierstrass-Mandelbrot fractal function, followed by performing micro-scale finite element (MFE) simulation analyses. MFE results reveal that assembly deformation primarily stems from global rigid-body displacements caused by elastoplastic deformations of surface asperities. Based on modified Hertzian contact theory, a surface contact potential energy model is established, enabling efficient deformation computation via swarm intelligence optimization. Comparative evaluations show that the proposed method achieves 90 % prediction accuracy relative to MFE benchmarks while reducing computational requirements by 99.8 %.</p>