Biomass estimation models (BEMs) employ power-law equations to estimate aboveground biomass (AGB) using basal diameter (BD) as the independent variable. Estimating AGB is crucial as regards calculating forest biomass and productivity on the ecosystem scale. One of the most widespread biomasses in dryland ecosystems is the Prosopis genus. However, most BEMs for Prosopis have been generated using relatively small datasets and have not been revised since their early inception. The objectives of this study were (1) to build a BEM for Prosopis pallida in order to estimate the AGB and tree biomass by fraction; and (2) to analyze the differences among the BEMs for the datasets of four priority Prosopis species (P. pallida, P. laevigata, P. glandulosa, and P. juliflora) so as to fit them into a single common model. This was done using both univariate and multivariate generalized linear models and five databases obtained from literature. Our results showed that the univariate and multivariate models had a high R2 and were able to predict AGB. Tree biomass by fraction was successfully modeled using both BD and tree height. The databases regarding natural and planted Prosopis forests were statistically different, signifying that they fit into two different equations.