Investment is a sum of money or other resources that are made at this time with the hope of obtaining benefits in the future. Stock is a common investment that is in great demand by investors. In investment activities there is the most important component, namely volatility. All financial evaluations require an accurate prediction of volatility. The volatility is identical to the conditional standard deviation of stock price returns. The important thing in investing apart from the rate of return is risk. Value at Risk (VaR) is a statistical risk measurement method that estimates the maximum possible loss on a capital market instrument at a certain level of confidence. To evaluate the quality of the VaR estimation, the model should always be backtested with the appropriate method. Backtesting is a statistical procedure in which actual gains and losses are systematically compared with the corresponding VaR estimates.  The purpose of this study is to determine the time series model to determine Value at Risk and backtesting. So that the mathematical model that will be used is the Autoregressive Moving Average-Glosten Jagannathan Runkle-Generalized Autoregressive Conditional Heteroscedastic (ARMA-GJR-GARCH) model. ARMA is a combination of AR and MA models, while GJR-GARCH is a development of the GARCH model by including leverage effects. The time series method used in this study is the ARMA-GJR-GARCH model in determining Value at Risk. Data processing is to determine the ARMA model, the GARCH model, then determine the GJR-GARCH model by looking at the heteroscedasticity and asymmetric effects on the GARCH model. Then the risk level of the ARMA-GJR-GARCH model that has been obtained will be calculated and the model will be validated by backtesting. The expected result is the best ARMA-GJR-GARCH model so that the model can be used to determine Value at Risk and obtain an accurate backtesting test. Based on the results of the study, it is hoped that investors can use it as a consideration in making investment decisions.