Shrimp farming is a key sector in economic development in Mekong Delta provinces. Unfortunately, there are many problems in shrimp farming, especially shrimp diseases which cause a considerable loss. Shrimp diseases are expressed through symptoms and manifestations of shrimp. Recognizing the importance of shrimp symptoms to help raise an early warning, in this research the authors apply several state-of-the-art text classification algorithms such as Logistic Regression, Random Forest, Naïve Bayes, Support Vector Machines, and Multilayer Perceptron on a collection of 1098 observations categorizing into 14 distinct classes. Several thorough evaluation scenarios have been conducted including a process tokenization and models’ comparison on the obtained data set with different ratios. The results show that Support Vector Machines achieves the highest classification accuracy (81.27%), followed by Multilayer Perceptron, Random Forest, Logistic Regression, and Naïve Bayes. Through the results of the study, it is feasible to apply machine learning algorithms to diagnose shrimp diseases entirely based on textual symptom descriptions.