Accurate forecasting of the reservoir inflow is crucial for operations and management of water resources. Due to the nonlinearity and nonstationarity of the real hydrological data, an empirical mode decomposition based long short term memory (EMD-LSTM) model is proposed in this paper for daily reservoir inflow forecasting upto 10 days lead time. The accuracy and performance of the model is analysed using the mean absolute error (MAE), root mean square error (RMSE) and Nash-Sutcliffe efficiency (NSE). The performance of the proposed EMD-LSTM model is compared with artificial neural network (ANN) and long short term memory (LSTM) model for 3 days, 7 days and 10 days ahead lead time respectively. Daily inflow data from 2013-2022 of the Bhakra reservoir located on river Sutlej in Himachal Pradesh, India is used to demonstrate the proposed model. The overall results of the model were highly encouraging in terms of having Nash-Sutcliffe efficiency of upto 0.94 in validation stage for 10 days ahead forecast as compared to the ANN and LSTM models. Therefore, the model can provide useful information when the models are used for decision making and can ensure safe operations of reservoir systems.