Because of all kinds of the difficulties to predict the future, agricultural business is a risky business with unstable incomes. However, farmers have a number of tools to manage such risky situations; for instance, they use diversification possibilities on farm contracts (e.g. multiple arable crops or insurance). The mix of activities and contracts should be balanced such that it provides the farmer protection and opportunities with respect to a wide range of contingencies. The portfolio modelling approach is often used to show how different combinations of activities and contracts stabilises farm income more than having a single activity without a contract. The conducted analysis in this thesis consists of four main steps. First, farm revenue variability was analysed between farms and within a farm. Within-farm analysis overviews the variability and dependency of crop revenue components (i.e., yields and prices) over time and their effect on farm revenue. Between-farm analysis refers to farm revenue variability that is affected by differences in business and financial characteristics of farms. Second, an evaluation of the dependency between expected gross margins and standard deviations was carried out using different optimisation programming. This analysis showed substantial differences between farms, which should be recognised in advising farms on portfolio selection. Then an analysis was carried out that clarifies the advantages and disadvantages of two portfolio optimisation approaches (parametric versus non-parametric). This analysis deals with the complex problem of specification and inclusion of risk into the analysis. Finally an evaluation of farm income stabilisation by diversification and insurance within farm portfolio context was presented. Three types of insurance products were considered: yield insurance, price insurance and revenue insurance. Insurance was found to be an efficient risk-management tool to stabilise farm income.