It is a difficult but critical task for financial analysts, investors, and researchers to forecast stock market prices. This article presents a unique method for predicting stock starting prices that employs Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) networks, both of which can detect patterns and temporal correlations in time series data. To prepare for modeling, we gathered historical stock price data, focusing on the opening, high, low, and closing values. We then normalized the data using MinMaxScaler. To avoid overfitting, we created a sequential model with LSTM layers and dropout regularization. 70% of the dataset was used for training, and 30% for model testing. This study lays a solid platform for financial analysts to make sound investment decisions while also adding to the body of knowledge by demonstrating the superiority of LSTMs and RNNs for stock price prediction. The LSTM model had an accuracy of 97.72%, precision of 97.98%, recall of 99.41%, and an F1 score of 0.9869, whereas the RNN model had 97.81%, precision of 98.26%, recall of 99.22%, and an F1 score of 0.9874. Future research will look into how to improve prediction accuracy even more by using variables like sentiment analysis and volume.