Résumé : An optimization procedure for sealing leakage paths in field-scale application of microbially induced calcite precipitation (MICP) is develop and applied to CO$_2$ storage. The procedure utilizes a recently developed field-scale MICP mathematical model implemented in the industry-standard simulator Open Porous Media (OPM) Flow. The optimization problem is solved using the ensemble-based optimization (EnOpt) algorithm where the objective function is defined such that maximizing calcite precipitation is done in the shortest possible MICP operational time. An injection strategy is developed to efficiently produce calcite in and around the leakage paths, and to avoid clogging unwanted areas of the reservoir, e.g., the near-well area. The injection strategy consists of combined injection of growth and cementation solutions in separate well segments to initiate the MICP process after establishing a biofilm in the leakage paths with an initial injection phase. The optimization procedure is applied to three synthetic CO$_2$ leakage scenarios. The numerical results show that the leakage paths in all scenarios are effectively sealed while keeping the total MICP operational time low.