The highly volatile price of Bitcoin makes it difficult for financial market players. This research aims to build a Bitcoin daily closing price prediction model using Long Short-Term Memory (LSTM) neural network. Bitcoin price data from January 1, 2014 to May 9, 2025 was taken from Yahoo Finance, normalized with MinMaxScaler, and divided into 80% training data and 20% testing data. The LSTM model, which consists of two LSTM layers (50 units each) and two dense layers, was trained with Adam optimization and mean squared error loss function. The model uses the 60-day price sequence to predict the next day's price. The evaluation results show high accuracy with Root Mean Squared Error (RMSE) 105.80, Mean Squared Error (MSE) 2,822,880.74, Mean Absolute Error (MAE) 1,103.42, and R-squared (R²) 0.995. This model becomes one of the reliable prediction tools for financial decisions using historical data. This research enriches machine learning-based bitcoin price prediction solutions.