Code search has received increasing attention in recent years due to the search for effective software development practices. However, traditional methods of code search often suffer from the problem of terminology inconsistency, where different search engines and code repositories may use different terminology. To address this challenge, researchers are suggesting a variety of ways for retrieving useful code snippets. This work proposes a unique model that incorporates BERT, Glove, LSTM and attention mechanism to successfully extract code snippets. LSTM is used to capture the sequential and contextual dependencies, whereas BERT and Glove are used to capture the contextual word representations. To assess the efficacy of the CSQE model, two benchmarked datasets are employed. The experimental findings show that the suggested model performs better in terms of assessment metrics then current baseline models.