Deep learning models have been widely adopted in stock price prediction due to their ability to capture complex temporal patterns in financial time series. Improving performance is an important issue to ensure the model can predict with high accuracy and reliability. In this study, we introduce an effective framework for stock price prediction, namely Hybrid CBLANet, which combines Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BLSTM), and an attention mechanism. We evaluated this model alongside other architectures including LSTM, BLSTM, CNN-LSTM, CNN-BLSTM, and their attention-enhanced variants on three benchmark datasets: AAPL, CSI300, and VNINDEX. Experimental results show that Hybrid CBLANet consistently achieves the best overall predictive performance, especially at the 14-days lookback window, given the lowest RMSE and MAPE, and the highest R \(^{2}\) across datasets. Thereby proving that our proposed Hybrid CBLANet is effective and stable across many datasets.