An adaptive dynamic search algorithm is proposed to predict the solubility of salt in a salt-water system. It is not subjected to the impact of the initial probe whether the results are convergent or not. The thermodynamic basis for the proposed algorithm is discussed in detail and the iterations of the algorithm and the calculation error are quantitatively studied. The multiple calculated results and the experimental solubility data are compared. The algorithm proves its high accuracy and efficiency in theory compared to other algorithms, such as the Newton iteration algorithm. It possesses reliable potential for application in low-salt system dissolution equilibrium calculations.