This paper presents a novel boundary-optimized fast Fourier extension algorithm for efficient approximation of non-periodic functions. The proposed methodology constructs periodic extensions through strategic utilization of boundary interval data, which is subsequently combined with original function samples to form an extended periodic representation. We develop a parameter optimization framework that preserves superalgebraic convergence while requiring only a few boundary node deployment, resulting in computatio…