Landslides are a major natural hazard that can cause significant damage and loss of life. They are often triggered by heavy rainfall, earthquakes, or other factors that can destabilize the soil and rock. To mitigate risks associated with landslides, it is important to predict where and when they are likely to occur. In this study, we developed a multimodel and multistep LSTM (MML) model for landslide prediction. The models were trained on historical weather and soil property data from Kamand Valley, Himachal Pradesh, India. Kamand Valley is particularly susceptible to landslides, and the data collected in this study provides a feasible resource for developing and testing landslide prediction models. The MML model is a novel approach to landslide prediction. The model was composed of two LSTM layers and was trained to predict the weather and occurrence of landslides at multiple time steps in the future. LSTM layers were used to learn the long-term temporal dependencies in the data. The model was evaluated using a variety of metrics, including mean absolute error (MAE) and F1 score. The results showed that the model could achieve a high accuracy in predicting landslide occurrence. The final model is deployed in a web service, where it can be used to make real-time predictions. The web service is designed to be user friendly and easy to use, and can be accessed by anyone with an internet connection.