In this paper, we have explored the potential of a non-stacked LSTM network combined with a Kalman Filter for denoising in the context of stock price prediction. The primary objective was to develop an optimal, non-ensemble model capable of accurately forecasting stock prices, which can be practically utilized for intraday trading as there is lack of model suitable for Intraday trading. Most models predicts stock direction either goes up/down and single price prediction close. We here predict all the prices open, high, low, close and volume which is helpful in intraday trading. Our approach leverages the inherent strengths of LSTM networks in handling finance time series data. The non-stacked LSTM model was chosen for its computational efficiency and for better prediction result obtained, making it a suitable option for real-time usage applications where speed and resource usage are critical factors. By integrating the Kalman Filter, we have enhanced the model’s ability to process noisy financial data, thereby improving the price prediction accuracy. In this we explores the deep learning models on Reliance dataset range from 2014-01-01 to 2024-07-07.