. One way reached is to determine mulberry of water volume in reservoir, especially to produce energy, by accomplishing optimization to manage reservoir. Difficult problem to utilize flow surplus and optimal to manage water reservoir is placed in uncapability to anticipate mulberry of water discharge in future. Remember such problem it can made prediction model of flow of water using “Flow of Water Rain Continue of Multiple Linear Regrestion model and “Discret Chain Markov Model”, so it can predict flow of water that enter Saguling Reservoir in next month based on data of hydrology and unit previously. Result from both model is examined to optimization model of Bellman Dynamic Program. In Chain Markov model is performed simplicity mulberry of water debit which enter to reservoir and divide it become 3 class. From result of computerization proses based on sensitivirty testing using Bellman Optimization model to synthesis flow of water by continue model of Linear Regression and Discret Chain Markov model, so the protected optimum route that using is the optimum route in extreme dry condition (Class 0), because by using such route can undertake critical reservoir condition (water deficit), whereas the optimal route in normal flow of water condition and extreme wet flow of water condition is not finish a critical reservoir condition.