Predicting cryptocurrency prices is difficult because financial markets are extremely volatile and non-stationary. Various deep learning techniques i.e. Transformers, Convolutional Neural Networks (CNNs), Long Short-Term Memory Networks (LSTMs), etc. have been used extensively for price prediction. However, these techniques are plagued by high computational overhead, over fitting, and poor long-range dependency in capturing performance. To improve the accuracy of crypto price prediction, this study proposes a new forecasting method based on the Mamba State-Space Model (Mamba-SSM) for the prediction of cryptocurrency prices. First, we use real-time crypto data from the Binance API. Then, we used three feature selection algorithms to select dozens of important features from hundreds of features in the crypto trading data. We formulate and test Mamba-SSM by utilizing historic price data along with technical signals to make future price predictions effectively. The proposed model is tested on Bitcoin (BTCUSDT), Ethereum (ETHUSDT), and Ripple (XRPUSDT) which show that Mamba-SSM model outperforms conventional models in terms of prediction effectiveness, stability, and computational optimization. This study identifies the potential of the Mamba-SSM as a cutting-edge tool for financial time series forecasting, providing enhanced predictive capability compared to existing deep learning models. The results indicate that the Mamba-SSM can be an effective tool for traders and investors to improve their decision making in cryptocurrency markets.