Construction and installation of foundation and anchorage elements often require large amounts of material and energy, which can lead to negative environmental and economic impacts. In contrast, even with material and energy constraints, tree root systems must ensure anchorage stability while simultaneously meeting other functions such as nutrient acquisition and water uptake. This study is focused on extracting tree root system-inspired anchorage principles and transferring them to the engineering application of anchors in dry sand. Architectural properties of a set of extracted orchard tree root specimens are used to develop L-system-based stochastic realizations. Eight bio-inspired architecture models spanning a spectrum of complexity ranging from fully simplified analogs to a natural root system are developed using L-systems via sequential restriction and adjustment of root architecture simulations. These models are 3D printed and tested in vertical pullout using centrifuge modeling to generate information on the features are most critical for development of efficient anchorage, such as peak capacity, stiffness, and softening. Projected area is determined to be the most important factor affecting peak capacity and residual capacity. Based on the complexity of the root models, the mechanical response of the simplest model is distinct, while the responses of the models with intermediate and high complexity can be classified into two different groups. Of the spectrum of designs tested, Models 5, 6, and 7 produce the most outstanding performances in terms of capacity per volume and prolonged post-peak residual capacity. The results shed light on the trade-off between capacity, stiffness, and efficiency in anchor design, and in challenges in construction and installation of tree root-inspired anchors.