Rainfall prediction is a noteworthy and challenging task in today’s world. Due to the nonlinear and dynamic behavior, advanced computer modeling is required to predict rainfall accurately, in lieu of unreliable statistical techniques. In recent days, Machine Learning (ML) such as Deep Learning (DL) and Artificial Neural Network (ANN) enables self-learning to build a data-driven module for a time-series dataset. This paper proposes a DL model called “Short-term Data-Driven Rainfall Forecasting Model” (SDDRFM) which uses Long Short-Term Memory (LSTM)-based rainfall prediction model for weather and rainfall forecasting in the region of Shillong, a city in Meghalaya. In this paper, a rainfall forecasting model SDDRFM is proposed with LSTM algorithm using the rainfall dataset of Shillong collected from NASA power grid data, and is compared with other ANN models such as Gated Recurrent Unit (GRU), Convolutional Neural Network (CNN), and Feedforward Neural Networks (FFNN). The model uses dominant features such as humidity, temperature, wind speed and has achieved an accuracy of 85.73% using LSTM which is better than other ML models.