The robust design method is applied to laser microdrilling to overcome the quality fluctuation caused by process uncertainty. Considering the stochastic uncertainty of noise factors (processing time, gas pressure, et al.) to affect the controllable pulse energy, pulse width and defocus in the laser hole-punching process, the exit diameter of hole is analytical robust optimal design. Firstly, through orthogonal experimental design, the quadratic regression equation between quality index and parameters is established by using response surface model. Secondly, the expectation and variance of index are obtained by using Taylor series expansion, and a robust optimal design model with minimum variance as the objective and expectation as the constraint is established. Finally, combining with optimization algorithm in MATLAB obtains the best combination of process parameters, and comparing with the deterministic optimization, which verifies the good effect of the robust method.