There exist complex competitive reactions among two acids and multiple minerals in mud acid acidification of sandstone reservoirs, which makes it difficult to accurately characterize the acid-rock reaction kinetics and provide basic data for the optimal design of sandstone acidification. To reveal the competitive mechanism of acid - rock reactions in sandstone reservoirs, six common sandstone mineral components were selected, and the dissolution rate was taken as the evaluation index. The laws of acid - rock reactions were analyzed through dissolution experiments and XRD diffraction, and the dissolution rates of multiple minerals in sandstone acidification were predicted by machine learning. The study shows that the sequence of dissolution rates of six common sandstone minerals is montmorillonite > kaolinite > chlorite ≈ illite > albite > quartz. When multiple minerals coexist, there is an obvious competitive mechanism among various minerals, which is quite different from the results of single minerals. Clay minerals have stronger competitiveness with acid fluid than quartz and feldspar. Among clay minerals, montmorillonite has the highest competitiveness, followed by kaolinite, and chlorite has higher competitiveness than illite. When albite coexists with chlorite or illite, it exhibits greater competitiveness in reacting with acid solutions compared to when it coexists with other clay minerals. Based on machine learning, a multi-layer perceptron neural network is used to develop a dissolution rate prediction method for sandstone acidification. The method achieves a prediction error of less than 20% and can accurately forecast the dissolution rates of multiple minerals. Research on the competitive mechanism of multi-mineral components and establishment of the dissolution rate prediction method can guide the optimization design of acidification technique for formations with different minerals.