Covid-19 has become a global health problem during this pandemic. Every country is struggling to fight this problem as well as Indonesia. Indonesia has a high number of new cases and this has an impact on the high demand for bed occupancy rates. To overcome this situation, we recommend the prediction of covid-19 using LGBM and LSTM. We implement two pre-processing, namely one-hot data encode and Normalization. The results of the pre-processing will be used as input for the prediction of new cases of COVID-19 using the LGBM and LSTM algorithms. The experimental results show that LSTM has better results than LGBM. We evaluated that the number of epochs we used in the LSTM had a large influence on the RMSE, MAE, and R2 measurements