An optimal cross-region scheduling of agricultural machinery can greatly contribute to minimizing the harvesting period in modern agriculture. However, the current management mode without a center cannot fully meet the flexible vehicle scheduling and highly efficient system, due to the different scheduling strategies and agricultural products among operating regions. Moreover, it is still lacking in the trust relationship between individual farmers and the cross-region dispatcher, due to the long distance between regions. It is a high demand to fully utilize the existing agricultural machinery resources, and effectively reduce the cost of purchasing agricultural machinery in each region. In this research, an intelligent scheduling system of agricultural machinery was constructed in cross-regional work using blockchain. A cross-region scheduling model of agricultural machinery was firstly established to evaluate the input cost under the number and demand of agricultural machinery, as well as the number of machinery operators. Then, the cross-region scheduling matching of agricultural machinery using the shortest distance was used to reduce the cross-region scheduling path cost, where the minimization of input cost was taken as the objective. The scheduling matching was automatically executed through the blockchain smart contract. A scheduling system was implemented in the form of decentralization to combine with the consensus algorithm of the blockchain system. The blockchain scheduling system was used to transfer the calculation process of the scheduling algorithm into each distributed user node in the system, where the main consensus was gained among all nodes. The scheduling matching results were then stored in the new block in real time as data storage after the consensus, together with the agricultural machinery and farmland information during the scheduling process. The reputation value was then rewarded for the user nodes that actively participated in the implementation. The nodes with the high reputation value were matched first in the future scheduling for the practical benefits. Finally, the field experiment was performed on the intelligent agriculture management platform, as well as the randomly generated scheduling tasks in the scheduling system. The cost and utilization of the scheduling within the region were compared using the farmland and agricultural machinery data of Yixing City, Jiangsu Province, and Guangde City, Anhui Province, China. The results show that the decentralized characteristics of blockchain fully met the needs of cross-region scheduling scenarios without a center. The trust was provided for the scheduling users using identity authentication, high transparency, and tamper-proof technical features of blockchain. The agricultural information was updated to deal with unexpected situations by transaction cancellation using reschedule matching and synchronized block. Once the total number of agricultural machinery was sufficient in each region, the cross-region scheduling reduced the purchase cost of agricultural machinery for resource utilization to 90%. Therefore, the input cost of cross-region scheduling decreased with the increase in the number of tasks, compared with the intra-region scheduling. Specifically, the input cost was reduced by about 21.3% when the number of tasks was 10. A better performance was achieved for the blockchain-based scheduling system. More than 800 scheduling transactions per second can be expected to process in the system when the number of user nodes and transactions in each block reached 16 and 800, respectively. The finding can provide a strong reference for the cross-region scheduling mode of agricultural machinery.