Large-scale natural disasters challenge the resilience of the surface transportation system. The objective of this research was to develop a resilience model of the surface transportation system with a mixed-traffic environment and considering varying Connected and Automated Vehicle (CAV) penetration scenarios. As deployment of CAVs is expected to improve traffic operations, a resilience model was developed in this research to evaluate the resilience performance of a transportation system with several CAV penetration levels (0%, 25%, 50%, 75%, and 100%) for a given budget and recovery time. The proposed resilience quantification model was applied on a roadway network considering several disaster scenarios. The network capacity in relation to trips at any phase of disaster was compared with the pre-disaster trips to determine the system resilience. The capacity variation and the travel time variation were also estimated. The analysis showed that the resilience of the transportation system improved with CAVs in relation to travel time and capacity improvement. Link travel times were significantly improved by higher CAV penetration rate. The findings also suggested that higher penetration of CAVs (i.e., 50% or more) increased the recovery costs. For example, the recovery costs needed for medium and large-scale disasters were 50% and 90% higher, respectively, compared with the recovery costs for a small-scale disaster. These higher costs were primarily for the repair and replacement of intelligent infrastructure required to support the operation of CAVs.