Abstract: How to effectively utilize sparse exploration data to accurately build geological models and optimize drilling plans step by step is of great significance in reducing exploration costs and improving the efficiency of collecting formation data. This article proposes an additional drilling optimization method based on the local coupled Markov chain (LCMC) model. This method establishes a geological model that adapts to complex and varying formations through fragmentary processing, local random modeling, and superposition of multiple fragments based on the drilling data of geological profiles, evaluates. The uncertainty of geological units is quantitatively evaluated using an entropy map generated by the LCMC model, and the optimal locations for additional drilling are predicted step by step based on the column average entropy curve. Research results show that compared to traditional methods, the proposed approach can achieve more rational and efficient drilling optimization, increase the accuracy of modeling complex formations with directional and inclination changes, and reduce the uncertainty of geological profiles.