Most microcapsule preparation methods produce a population of microcapsules in a bulk solution. To control the microcapsule preparation or obtain an optimal preparation condition, the mechanism of the microcapsule preparation should be investigated. The mechanism is estimated via structure reformation during the preparation process because diameter and wall thickness are drastically altered in the solution. Considering microcapsule applications, some important properties, such as the mechanical properties of microcapsules and release rate of the encapsulated product, depend on the microcapsule structure. In this study, polystyrene microcapsules containing saline water droplets were prepared via the solvent evaporation method from a solid-in-oil-in-water (S/O/W) emulsion system. The microcapsules exhibited a speci c structural distribution, which comprised monocore, multicore, and solidcore structures. The structural distribution was altered by the preparation condition. The monocore structure was absolutely dominant owing to the increase in the amount of calcium chloride added in the organic phase. The salt concentration is not the sole controlling factor of the microcapsule structure, as the surfactant and dispersion exerted a signi cant impact on the microcapsule structure. The structural distribution was automatically analyzed by a machine learning algorithm (MLA). The decision-making time for the microcapsules preparation was shortened by the accelerated structure determination, and the accuracy was improved by increasing the number of counting particles.