The article presents an approach to modeling risk propagation in multi-station systems using Monte Carlo simulation. It is assumed that the risk at each workstation is a random variable with a prescribed initial distribution, and that its evolution depends on two key mechanisms: retention (internal holding of risk) and transfer (flow of risk between stations). The model is built on an adjacency matrix describing the network of connections among workstations and on a recursive evolution equation that includes a stochastic term to capture unpredictable fluctuations. Monte Carlo simulations were carried out in three variants: baseline (standard parameter values), enhanced-transfer (increased risk-flow coefficients) and heightened-retention (increased capacity to hold risk). Results were analyzed from two viewpoints: statistical (mean, variance, skewness and kurtosis of the total system risk after T simulation steps) and temporal (the evolution of risk over successive stages). Findings are illustrated with risk-distribution histograms and time-series plots. The model was validated on a case study of a window-frame manufacturing process, in which each technological operation was treated as a distinct risk source, able to retain or transfer risk to subsequent operations. The analysis revealed significant cascade effects: under the enhanced-transfer scenario, risk spreads rapidly through the system, whereas with higher retention it accumulates locally. These insights enable identification of critical workstations and optimization of mitigation strategies, thereby supporting the design of more resilient production systems.