E-students' satisfaction significantly affects the success of e-learning adoption and leads to quality improvement of e-courses. E-students' satisfaction also highly impacts student learning performance and learning activities in e-courses. However, many previous studies focused on e-student satisfaction factors in developed countries and in some big cities in Vietnam in normal educational contexts. Limited research emphasized top-ranking factors for e-students' satisfaction in emergency remote learning (ERL) in universities in rural regions in Mekong Delta using machine learning methods. Thus, this study aimed to identify crucial factors affecting e-students' satisfaction in ERL in a public university in rural areas of Mekong Delta in Vietnam by using Random Forests (RF) and Support Vector Machine (SVM) algorithms. This study used a 5-point Likert e-survey including 28 items allocated to four scales: 1) Course designs, (2) Infrastructure and technology, (3) Teacher-student interaction; and (4) Lecturers to collect research data. The employed data included responses of 916 e-students from multi departments of the target university during the second semester of the 2021/2022 academic year. The results displayed that RF (94.47%) outperformed SVM (91.15%). The crucial factors were mainly related to infrastructure and technology, materials, and teacher-student interaction, especially feedback and evaluation. The outcomes of this study provide sustainable e-learning in the post-pandemic educational context. The proposed prediction models inform policy-makers and e-learning program designers in a long-term strategy that meets e-leaners' needs in order to attract more e-students and maintain e-student retention.