One of risk measurements – Value-at-Risk (VaR) was analyzed in this work. Complete definition of VaR is presented and three classical calculation methods of it are examined: parametric, historical simulations, and Monte-Carlo generations. The main advantages and disadvantages of the application of VaR are reviewed. Correlation effect for two assets risk diversification is examined and Markowitz method for calculation of efficient frontier is presented. Analyzed methods were implemented in the program, which calculates the first moments of portfolio’s returns, correlations between different assets and for a given return adjust weights in a manner to minimize dispersion of portfolio’s returns. For every efficient portfolio VaR can be calculated at any confidence level. Created program was used to analyze three investment portfolios: one of generated data with normal distribution, one of LITIN-10 index stocks and one of OMX Vilnius index stocks. Efficient frontier for these portfolios and VaR for whole efficient frontier were calculated. We noticed difference between minimal VaR and minimal standard deviation portfolios, consequently three investment strategies were implemented. The best results for analyzed portfolios there achieved with minimized VaR strategy.