This study analyzes the interrelationships between weather parameters, including average temperature (Tavg), relative humidity (RH_avg), rainfall (RR), and average wind speed (ff_avg) in North Jakarta and Central Jakarta, and compares the performance of Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) models in predicting these parameters. Data was collected from Tanjung Priok Maritime Meteorological Station in North Jakarta and Kemayoran Meteorological Station in Central Jakarta from December 2021 to December 2024. The results show that GRU performs better in North Jakarta, with RMSE of 9,02, MSE of 81,28, and MAE of 4,21 at 75 epochs, while LSTM yields RMSE of 10,02, MSE of 100,34, and MAE of 4,62 at 50 epochs. Conversely, LSTM outperforms GRU in Central Jakarta, with RMSE of 8,96, MSE of 80,22, and MAE of 4,65 at 100 epochs, while GRU produces RMSE of 9,53, MSE of 90,78, and MAE of 4,85 at 75 epochs. GRU is more effective in capturing extreme fluctuations, while LSTM excels in predicting interrelationships between parameters. This study provides insights into selecting the appropriate weather prediction model based on the priority of prediction accuracy or the ability to capture extreme fluctuations