Planning for NASA's Asteroid Redirect Mission (ARM) requires estimating the forces that appear during extraction of a boulder from the surface of an asteroid with unknown surface regolith properties. These forces are estimated for a vertical constant force or acceleration pull and a rolling, constant force, torque (peel) on a 4-m diameter spherical boulder using both analytic and discrete element method (DEM) models considering the effects of microgravity and regolith cohesion using Johnson–Kendall–Roberts (JKR) model. Estimates of the bulk asteroid regolith cohesion strength derived from lunar and asteroid regolith studies ranged from 25Pa to 250Pa. JKR cohesive forces at particle contacts depend on particle surface energy and effective curvature radius (particle size). DEM particle size dependent cohesion parameters are linked to estimated regolith cohesion strength by simulating shear and tension tests over a range of DEM particle surface energies resulting in the formulation of the dependence of particle surface energy as a function of cohesion strength and particle size. Maximum extraction forces occur for a vertical pull through the boulder center of mass with constant acceleration. Extraction force decreases for a constant force pull to 0.62pcS where S is the boulder surface area embedded in the regolith and pc is the cohesion strength of the regolith. Boulder extraction by peeling produces the smallest forces by up to more than a factor of 2, as the failure across the boulder surface increases progressively rather than being fully engaged as occurs during a vertical pull extraction. Variations between DEM and analytic results differed from 9% to 17% over the range of regolith cohesion values and peel extraction leverage.