Bitcoin and other cryptocurrencies are technological innovations that have transformed the world of finance. However, these developments introduce new risks. Because of the relative anonymity these systems offer, money laundering is one of the most significant risks connected to cryptocurrencies. This anonymity makes it more difficult to identify money obtained through illegal means, which in turn makes it possible for criminal operations to continue and grow. As a result, creating efficient techniques to identify and stop money laundering in cryptocurrency transactions has grown in importance as a research challenge. In this study, we suggest a hybrid model based on deep learning to detect illicit money transfers in cryptocurrency transactions. Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory Networks (BiLSTM), and an Attention Mechanism are all integrated in the suggested method. CNNs are used in the model’s initial stage to extract significant features from the unprocessed data. The BiLSTM layer then received these features in order to identify the dependencies in the sequential data structures. The Attention Mechanism enhances the overall classification performance in the last step by giving the BiLSTM outputs importance weights. The Elliptic dataset was used to assess the suggested model’s performance. The experimental results indicate that the model achieves superior performance compared to existing methods, with an accuracy of 97.9%, precision of 99.0%, recall of 98.0%, and F1-score of 98.0%. The findings of this study highlight the effectiveness of deep learning models enriched with Attention Mechanism for detecting illicit activities in cryptocurrency transactions. Beyond presenting a high performing model, its contribution to the literature lies in offering an novel approach to prevent crimes associated with the growing use of cryptocurrencies.