Amino acids are one of the main building blocks of proteins, and because of this connection and their main role in biochemistry, they are of great interest in pharmaceutical industries today. In nature, there are approximately 300 amino acids, out of which 20 are present in the human body and are responsible for protein synthesis. As a result, amino acids are widely used in various fields such as biological sciences, chemistry, medicine, and various industries. This research focuses on estimating the acid dissociation constant (pKa) of the amino group related to 52 amino acids using the quantitative structure–property relationship (QSPR) method, employing four different regression models: Genetic Algorithm-Multiple Linear Regression (GA-MLR), Particle Swarm Optimization-Support Vector Machine (PSO-SVM), Feedforward Neural Network (FFNN), and Decision Tree (DT). Using the GA-MLR, the optimal number of descriptors for estimating pKa was determined to be 9. By comparing different machine learning models, it can be observed that the coefficient of determination (R2) follows the trend R2 PSO-SVM = 0.9681 > R2 FFNN = 0.9539 > R2 DT = 0.9329 > R2 GA-MLR = 0.9231. Therefore, the PSO-SVM model provided the best prediction, while the FFNN, DT, and GA-MLR models ranked next in terms of accuracy in predicting the pKa of amino acids. By dividing the data into test/train sets in a ratio of approximately 20/80, the model's validity was confirmed due to the nearness of the R2 values between the training and testing sets. Ultimately, the best model is capable of predicting the pKa value of the amino group in amino acids with a mean absolute percent error of 1.18 %.