Current technology has limited applicability for electroporation based treatment of deep-seated tumors, and is in general, not optimized in terms of compliance with clinically relevant parameters. Here we present a novel electrode device developed for electrotransfer of antineoplastic drugs and genes to intracranial tumors in humans, and demonstrate a method to optimize the design (i.e. geometry) of the electrode device prototype to improve both clinical performance and geometrical tolerance (robustness). We have employed a semiempirical objective function based on constraints similar to those used in radiation oncology. The results show that small geometrical changes may yield a significant improvement. For example, a 2. mm displacement of 6 electrodes yields 14% better compliance with the clinical parameters, compared to the prototype, and additionally makes the electrode device less sensitive to random geometrical deviations. The method is readily applicable to other electrode configurations.