<p>The COVID-19 pandemic affected the global education system, leading to the sudden closure of institutions and the subsequent adoption of e-learning. The limited e-learning infrastructure among instructors and students in developing countries such as Bangladesh hindered this transition during the COVID-19 crisis. The primary objective of our study is to assess the factors influencing preference for e-learning over traditional learning during the COVID-19 pandemic, as well as the depression and anxiety associated with e-learning among Khulna University students. Using a two-stage non-probability sampling approach involving purposive and snowball recruitment and validated scales to measure anxiety (GAD-7) and depression (PHQ-9), a cross-sectional survey was conducted among 500 students. The study revealed that preference for e-learning over traditional learning was associated with students living in urban areas (AOR: 2.87; 95% CI: 1.26–6.57) and those with good internet speed (AOR: 3.97; 95% CI: 1.78–8.87). In contrast, lower preference for e-learning was observed among students who reported inadequate online class support, difficulty handling devices, and lower perceived teacher engagement. Students who reported online class difficulties, perceived monotony, technical problems, and lower preference for e-learning had higher odds of reporting anxiety and depression symptoms. Strengthening e-learning infrastructure and providing adequate training and support for both students and teachers are crucial, along with robust support for students to ensure educational continuity during future crises.</p>