This paper presents a probabilistic model for forecasting the global spread of infectious diseases, using a combination of local infection dynamics and stochastic transport on a worldwide aviation network. The model incorporates the heterogeneity of the global aviation network and stochastic fluctuations in disease transmission, recovery, and dispersal. It is applied to the SARS outbreak and shows good agreement with published case reports. The model demonstrates that the high predictability of epidemic spread is due to the network's heterogeneity. The model can be used to predict future epidemic spread and identify high-risk regions. It also analyzes the effectiveness of different control strategies, showing that rapid and focused responses are essential to prevent global spread. The model accounts for the stochastic nature of disease transmission and recovery, and uses a stochastic SIR model with a latent stage. The model is tested on the SARS outbreak and shows good agreement with real-world data. The results show that the model can accurately predict the spread of SARS and that the impact of fluctuations is less significant than expected. The model also demonstrates that the predictability of epidemic spread is enhanced in networks with high variability in transition rates. The model is used to evaluate control strategies, showing that isolating a small fraction of the largest cities can significantly reduce the need for global vaccination. The study highlights the importance of understanding the global aviation network in developing effective control strategies for infectious diseases.This paper presents a probabilistic model for forecasting the global spread of infectious diseases, using a combination of local infection dynamics and stochastic transport on a worldwide aviation network. The model incorporates the heterogeneity of the global aviation network and stochastic fluctuations in disease transmission, recovery, and dispersal. It is applied to the SARS outbreak and shows good agreement with published case reports. The model demonstrates that the high predictability of epidemic spread is due to the network's heterogeneity. The model can be used to predict future epidemic spread and identify high-risk regions. It also analyzes the effectiveness of different control strategies, showing that rapid and focused responses are essential to prevent global spread. The model accounts for the stochastic nature of disease transmission and recovery, and uses a stochastic SIR model with a latent stage. The model is tested on the SARS outbreak and shows good agreement with real-world data. The results show that the model can accurately predict the spread of SARS and that the impact of fluctuations is less significant than expected. The model also demonstrates that the predictability of epidemic spread is enhanced in networks with high variability in transition rates. The model is used to evaluate control strategies, showing that isolating a small fraction of the largest cities can significantly reduce the need for global vaccination. The study highlights the importance of understanding the global aviation network in developing effective control strategies for infectious diseases.