<p>In this work, a student project report retrieval system is designed that identifies the relevance of the student queries with appropriate project domains and retrieve the most relevant reports quickly. To design this model, a project report repository is created collecting the student project reports of five years of an Institute. A Deep Neural Network (DNN)-based domain classification technique (DCT) is applied for finding the relevance of the student queries with respective domains for faster project report retrieval. A query expansion technique is also applied to address all possible word variations that may convey the student’s intent to ensure retrieval of all relevant reports from the repository. To rank the retrieved project reports as per their relevance to the student’s queries, a Query-Report Relevance (QRR) score-based ranking algorithm is applied. Several experiments were performed to analyze the performance of the presented model. The accuracy of the DNN-DCT is evaluated with different student queries with average classification time of 0.3&#xa0;s and average domain-wise accuracy of 90%. The precision and recall measures of the retrieval model on varied student queries are also evaluated with respect to different domains. The average scores obtained for the same are 96% and 95% respectively in the considered domains. The traditional Boolean Model (BM) and the Fuzzy Clustering-based Semantic Retrieval (FCSR) model are considered to compare the performance of the presented model in project report retrieval for different domains. The presented model achieves 10% increased accuracy percentage over the two traditional models showing the effectiveness of the used methodology in the project report retrieval system.</p>