In Indian agriculture, production of rice crop occupies 40% of total crop production. Rice is cultivated in wide range of location and under a variety of climatic conditions. Predicting crop yield in different climatic condition is highly intensive which improve the farmers ability in decision making of choice of crop. This study aims to do prediction model with artificial neural network. five different training algorithms of neural network for prediction are studied here along with different feature selection methods. Feature selection is very important to make better prediction hence suitable feature selection methods was identified and then continued with the prediction model. Data is obtained from publicly available Indian data from 1961 to 2014.The parameters considered here were total area harvested, precipitation, rainfall, min,max and average temperature, Emission of CH4, reference crop evapotranspiration, Emission of Co2, Area under irrigation and yield. Here multilayer neural network model is constructed and visualized using R and compared with the linear model. © 2017, Institute of Advanced Scientific Research, Inc. All rights reserved.