Declining sea ice in the Arctic Ocean is exposing its coasts to more frequent and intense forms of wave energy and storm surge. As a result, erosion rates along some stretches of coastline in the Alaskan Arctic have doubled since the middle of the 20th century and now rank among the highest in the world. People concentrated near the coast are being heavily impacted by erosion, with some facing relocation. Coastal erosion is projected to increase the cost of maintaining infrastructure by billions of dollars in the coming decades. The financial impact of enhanced erosion will likely be further exacerbated by emerging geopolitical pressures, including the discovery of natural resources, opening of new shipping routes, and construction of support facilities in the Arctic. Scientific knowledge and engineering tools for predicting coastal erosion and guiding land-use decision are not well-suited for the ice-bonded bluffs of the Alaskan Arctic. Investigation of the oceanographic, thermal, and mechanical processes that are relevant to permafrost bluff failure along Arctic coastlines is thus needed. We introduce a geomechanical simulation framework, informed by field observation and laboratory testing, that focuses on the impact of bluff geometry and material variability on permafrost bluff stress states associated with a 9-km stretch of Alaskan Arctic coastline fronting the Beaufort Sea that is prone to toppling-mode block failure. Our approach is advantageous in that it is based on measurable physical properties (e.g., the bluff geometry, permafrost bulk density, Young’s Modulus, and Poisson’s Ratio) and does not require the potential failure to be defined a priori, but rather, the failure area can be interpreted from the multidimensional patterns of stress produced by the model. Our findings highlight how (1) block failure characteristics could be tied to variations in the intensity and duration of the storm energy that intersects the coastline and (2) how deformation processes that create non-uniform patterns of displacement may play a role in localizing block failure. We propose that this kind of physics-based simulation approach can facilitate hypothesis testing regarding the prediction of decadal-scale erosion rates for increasingly dynamic coastal permafrost systems.