With the increased popularity of the Internet for instructional purposes in the recent years, learners can now ably build a positive attitude towards its use in their learning activities through constant interaction with various information technologies. The purpose of this study was three-fold: to examine the extent to which students’ self-reported perceptions of Internet use constituted meaningful and interpretable dimensions of the Internet use in learning construct; to test the construct adequacy in terms of its validity and reliability; and establish if the measurement model is an adequate measure of Internet use for learning across levels of internet knowledge. The sample consisted of 258 undergraduates from six faculties of Islamic University in Uganda. The data were collected using a self-reported 16-item questionnaire that measured undergraduate students’ perceptions on Internet use for learning and analysed using Structural Equation Modeling and Confirmatory Factor Analysis with AMOS software version 22.0. The study findings reveal that students’ Internet use for learning is a valid and reliable multidimensional construct with AVEs>0.5, shared variance <0.5 and composite and Cronbach Alpha >0.7; and the measurement is Internet knowledge-invariant with chi-square change p>05. The results supported and extended the results of previous work on students’ perceptions about internet use for learning. The findings are useful for evidence-based decision making regarding the integration of internet applications in student learning activities, and in enhancing future research in the domain of Internet use and pedagogy.