As the markets for cryptocurrencies have risen rapidly in recent years, there is more interest in forecasting their prices. The use of Long Short-Term Memory (LSTM) neural networks for cryptocurrency price prediction is examined in this paper. The model is trained to predict the price using historical data on the cryptocurrency exchange rate with the currencies of G20 nations. The pre-processed data collection is divided into training and testing sets. Then, using the training set and testing set, the LSTM model is trained. To assess the model's performance, it is put up against the time series model known as Autoregressive Integrated Moving Average (ARIMA). The outcomes show that the LSTM model performs better in terms of accuracy and offers more trustworthy forecasts of the cryptocurrency exchange rate than the ARIMA model. These results imply that LSTM neural networks are a potential method for predicting the price of cryptocurrencies and may be used to help traders and investors make wise choices.