This research developed reliable prediction models for the unconfined compressive strength (UCS) of soil stabilization with common cementitious binders, such as lime, cement, fly ash, and blast-furnace slag. Several Artificial Intelligent (AI) techniques, including Artificial neural network (ANN), Multi-Gene Genetic Programming (MGGP), and Gene-expression programming (GEP), are applied to generate the prediction formulas. Some experimental data from the doctoral thesis of Dr. Bolton and Dr. Do are used to develop AI-based models as case studies. In addition, a thousand comprehensive data points gathered from several experimental studies in the literature are used for the model development. The soil characteristics, the making method, the curing period, the binder types, and binder contents are all considered as the independent variables in the models. The research results show that the proposed AI-based models perform well with a high correlation coefficient and low errors (e.g., RMSE and MAE). Hence, these formulas could be confidently applied in estimating the UCS of soil stabilization with different binders. Furthermore, a comparative study is conducted to evaluate the predictive ability and the performance of the ANN, MGGP, and GEP models. Besides, parametric studies and sensitivity analyses are carried out to examine the effects of the essential parameters on the UCS of stabilized soil. The research findings could help engineers choose suitable binder types and cost-effective methods to optimize the UCS of stabilized soil.