Cardiac resynchronization therapy (CRT) has shown to improve hemodynamics and clinical symptoms of congestive heart failure. The present article investigates an automated non-invasive strategy based on a computer model of the heart to optimize biventricular pacing as a CRT with respect to electrode positioning and timing delays. Accurate simulations of the electrical activities of the heart require suitable anatomical and electrophysiological models. The anatomical model used in this work, is based on segmented MR data of a patient in which a variety of tissue classes for left ventricle are considered based on AHA standard in accordance with fiber orientation. The excitation propagation is simulated with the ten Tusscher et al. electrophysiological cell model using an adaptive cellular automaton. The simulated activation times of different myocytes in the healthy and diseased heart model are compared in terms of root mean square error (ERMS). The results of our investigation demonstrate that the efficacy of biventricular pacing can greatly be improved by proper electrode positioning and optimized A-V and V-V delay.