This paper proposes a deep learning model that forecasts waste by using the Long Short Term Memory (LSTM) technique to optimize all aspects of the process for waste management. The model was able to answer all questions related to the role that environmental and geospatial factors play in predicting waste. We utilized a waste dataset spanning over 5 years that includes geospatial and environmental factors. The training data was prepared using a sliding window technique and fed into the LSTM model. The model is trained to predict future waste generation based on past trends and geographical data such as population, climate, etc. Performance metrics including Mean Squared Error (MSE), Relative Absolute Error (RAE), and Relative Squared Error (RSE) have been used for testing the performance of the proposed model and showed comparable results. Thus, the outputs of the proposed model provide valuable insight for smart cities, aid in the long-term planning of the cities, and improve the schedules for resource distribution and waste collection.