Whenever an investor decides about his wealth allocation, investment return (expected return) and risk are considered. In financial practice the concept of forming an optimal portfolio has become popular in the last 50 years and this thesis researches how to form one with combination of two or three risky assets traded on German market. Specifically, the forming of optimal portfolios is done with combination of three securities: BASF, Deutsche Bank and Siemens and the market index DAX serve as market portfolio. Analysis is made with respect to various types of investors, primarily according to their risk aversion. The classical part of portfolio analysis relies on Modern Portfolio Theory and measures risk through volatility (variance and covariance) and implies Sharpe ratio as a portfolio indicator. Afterwards, application of real data in common financial models, such as CAPM and VaR, is compared to reality. Advanced part of portfolio analysis introduces the concept of semi deviance (downside deviation) and contains Sortino ratio and forecasting method EWMA with different Lambda values. The research ends with computing the correlation between the two known portfolios without knowing their covariance. Main purpose of the thesis is to define several possibilities of wealth allocation and to form a portfolio of the analyzed securities as well as to apply obtained data into developed financial models in order to determine how well such models estimate real data.