Bitcoin’s recent volatility has captured significant attention, prompting extensive research into the factors driving its price fluctuations. This study presents a comparative analysis of seven machine learning models—Decision Tree, Random Forest, XGBoost, K-Nearest Neighbors (KNN), Support Vector Regression (SVR), Artificial Neural Networks (ANN), and Long Short-Term Memory (LSTM)—for forecasting Bitcoin prices using data from 2016 to 2023. The models were evaluated over forecasting horizons of 7, 30, and 90 days. The results demonstrate that deep learning models, particularly ANN, consistently outperform traditional models. For the 7th-day forecast, the ANN model achieved an RMSE of 0.0722, MAE of 0.0542, and MSE of 0.0052, significantly outperforming traditional models with RMSE values around 1.4. Similar trends were observed for the 30th and 90th-day forecasts, where ANN and LSTM models showed superior accuracy with lower error rates compared to traditional models. SVR consistently underperformed across all forecasting periods, showing the highest error rates. The study highlights the effectiveness of deep learning models in capturing the complex, non-linear patterns of Bitcoin price fluctuations, providing valuable insights for financial forecasting in the cryptocurrency market.