This study developed, analyzed, and compared mission architectures for human exploration of Mars' moons within the context of an Evolvable Mars Campaign. METHODS: All trades assumed conjunction class missions to Phobos (approximately 500 days in Mars system) as it was considered the driving case for the transportation architecture. All architectures assumed that the Mars transit habitat would remain in a high-Mars orbit (HMO) with crewmembers transferring between HMO and Phobos in a small crew taxi vehicle. A reference science/exploration program was developed including performance of a standard set of tasks at 55 locations on the Phobos surface. Detailed EVA timelines were developed using realistic flight rules to accomplish the reference science tasks using exploration systems ranging from jetpacks to multi-person pressurized excursion vehicles combined with Phobos surface and orbital (L1, L4/L5, 20 km distant-retrograde-orbit [DRO]) habitat options. Detailed models of propellant mass, crew time, science productivity, radiation exposure, systems and consumables masses, and other figures of merit were integrated to enable quantitative comparison of different architectural options. Options for prestaging assets using solar electric propulsion versus delivering all systems with the crew were also evaluated. Seven discrete mission architectures were evaluated. RESULTS: The driving consideration for habitat location (Phobos surface versus orbital) was radiation exposure, with an estimated reduction in cumulative mission radiation exposure of up to 34% (versus a Mars orbital mission) when the habitat is located on the Phobos surface, compared with only 3% to 6% reduction for a habitat in a 20-km DRO. The exploration utility of lightweight unpressurized excursion vehicles was limited by the need to remain within 20 minutes of solar particle event radiation protection combined with complex guidance, navigation, and control systems required by the nonintuitive and highly-variable gravitational environment. Two-person pressurized excursion vehicles as well as mobile surface habitats offer significant exploration capability and operational benefits compared with unpressurized extravehicular activity (EVA) mobility systems at the cost of increased system and propellant mass. Mechanical surface translation modes (ie, hopping) were modeled and offered potentially significant propellant savings and the possibility of extended exploration operations between crewed missions. Options for extending the use of the crew taxi vehicle were examined, including use as an exploration asset for Phobos surface exploration (when combined with an alternate mobility system) and as an EVA platform, both on Phobos and for contingency EVA on the Mars transit habitat. CONCLUSIONS: Human exploration of Phobos offers a scientifically meaningful first step towards human Mars surface missions that develops and validates transportation, habitation, and exploration systems and operations in advance of the Mars landing systems.