Abstract: Considering the endogenous and exogenous uncertainty problems of logistics service supply chain, based on complex science management qualitative analysis tools-Exploring Graph, risk early-warning index system of logistics service supply chain is constructed. Then based on the theory of entropy, index weights are calculated by using "Structure Entropy Weight Method". Next, on the basis of them, Markov chain forecasting method is integrated into the process of fuzzy comprehensive evaluation, dynamic risk warning evaluation model of logistics service supply chain is constructed. Finally, the effectiveness of the model is validated through an example, it provides a new way for early-warning evaluation of the logistics service supply chain.