Based on a previously developed and verified helical spring dynamic model, optimization and FFT routines have been added to the spring dynamic model as controlling and postprocessing units to form a computer-based optimal design tool. The objective function minimized is the amplitude of spring resonance. Design variables are the coefficients of a polynomial which describes the variation of the diameter of spring wire along the spring helix. Constraints are the maximum fatigue stress on any location of spring wire, the minimum and maximum wire diameter, and the maximum required spring force. In a case study for an automobile engine valve spring, resonant harmonic power has been reduced by 47%, given the newly designed spring a potential to be used at 8.3% higher engine speed before potential valve toss occurs.