Abstract: The SICOMINES copper cobalt deposit is located in the south west side of Kolwezi, Congo (Kinshasa) and is a typical deposit in the Katanga copper belt in Central Africa. Because of the complex genesis, such as tectonic hydrothermal superimposition and oxidation leaching of the deposit, there are more than dozens of copper cobalt minerals in the ore and their processing properties varies greatly. That caused some issues in the mine, such as complex process, unstable production and low recovery efficiency of Co. In this paper, the metallurgical geology modeling of cobalt was conducted by using Datamine and Leapfrog software for the first time to improve these issues. Firstly, historical exploration data were collected, strata and mineralized domain models were established, and the variation rule of cobalt grade in space was preliminarily obtained. Secondly, sampling design was carried out to collect process mineralogical samples that represent the grade distribution of cobalt in strata and ore bodies. Thirdly, quantitative data of mineral content and cobalt occurrence state in each sample were obtained by process mineralogical method, and these data were written into the model by using interpolation methods such as single domain assignment method and distance inverse power ratio method. Then, according to the spatial distribution of cobalt mineral beneficiation types, five spatial beneficiation zones were divided, namely suitable for flotation (TYPE1), suitable for magnetic separation (TYPE2), suitable for magnetic separation + flotation (TYPE3), suitable for leaching (TYPE4) and difficult to recovery (TYPE5), and a preliminary geometallurgical model was constructed. Finally, comprehensive samples were collected from the 5 beneficiation fields respectively for beneficiation experiments. The experimental results showed that, by using the beneficiation process used in mine production, the difference of cobalt recovery efficiency and concentrate grade in the 5 beneficiation fields was obvious. The original production process was only applicable to the spatial domain TYPE1, TYPE2 and TYPE3, indicating that the classification of beneficiation fields in the geometallurgical model was basically reasonable. The geometallurgical model provides guidance for realizing classification mining - classification sorting according to different properties of cobalt minerals, and improving cobalt recovery and production stability.