Value at Risk (VaR) has become a standard measure of market risk employed by many financial institutions for both internal and regulatory purposes. Measuring risk is a classical problem in Statistics, Economics and Finance. The regulators and financial executives always pursue the effective way to measuring the risk. Some VaR-like concepts have been founded through retrospective analysis in this history. VaR is defined as a distinct concept in late 1980s, triggered by the stock market crash of 1987, the first major financial crisis. VaR is defined as the value that a portfolio will lose with a given level of confidence over a certain time horizon. In this thesis, we calculate VaR under the conditions of without jumps and with jumps. We use the high-frequency financial data of Shenzhen Composite Index to make a real analysis, and researches show that it is more accurate to compute VaR by using high-frequency financial data than low-frequency financial data.