The present study aims to forecast rainfall intensity on a daily basis which could be used as an input to a rainfall threshold based landslide early warning system. The study area is Coonoor, a potential site of landslides, located in Nilgiris district of Tamil Nadu, India. Artificial neural network (ANN) and extreme learning machine (ELM) models are developed to forecast rainfall intensity one day in advance and their performance is compared. The input data set for the predicting models comprises of rainfall intensity, minimum and maximum temperature, cloud cover, wind speed and relative humidity were recorded on a daily basis at Coonoor observatory for a period of 10 years. The study also computes the risk of landslide occurrence by comparing the predicted rainfall intensity with the rainfall threshold for the study area. The performance of the forecasting models was evaluated using standard performance evaluation measures-mean square error (MSE), correlation coefficient (CC) and time taken for convergence. The results obtained confirm the suitability of both ANN and ELM models for the task of rainfall forecasting applied to landslide prediction.