The paper focuses on practices of business valuation, in which habitually and most frequently only point estimates of inputs are regarded. It shows that under many common situations this process may be misleading and incorrect. It emphasizes that the pure deterministic analysis relying on point estimates ignores valuable information related to an uncertainty of the authors' estimates. As a solution this paper suggests that the model should be based on probability distributions and the Monte Carlo Simulation, which enables incorporating the risk and uncertainty into the business value estimate. The simulation approach also offers relatively easy solutions to common problem areas, including the integration of expert opinions, nonrecurring events and dependencies between model variables.