During the deployment of quantum key distributed optical networks, limited quantum key resources, uneven distribution of network resources, and chaotic scheduling can result in large-scale business congestion and seriously impact optical network security. To overcome this issue, a quantum key resource allocation scheme is proposed. First, a linear programming constraint model of minimum wavelength and time slot consumption is constructed to reduce resource allocation. Second, considering the security requirements of a business, a key resource allocation method based on business priorities is proposed, combined with a dynamic key update mechanism and flexible security level adjustment strategy. Finally, a routing, wavelength, and time slot allocation algorithm is proposed, using resource evaluation values as routing evaluation indicators to identify paths with adequate resources and fewer hops while enhancing time slot utilization by establishing dynamic threshold values. The simulation results reveal that the proposed scheme can efficiently and uniformly use the resources in the network. Compared with the NPFSL-RWTA and SSL-RWTA algorithms, the success rate of optical path connection requests in the NSFNET (UBN24) network is increased by 7.9% and 16.3% (4.8% and 20.9%), respectively, considerably reducing the occurrence of business congestion and improving the optical network security.