Linkage disequilibrium (LD) measured over the genomes of a species can provide important indications for how future association analyses should proceed. This information can be advantageous especially for slow-growing, perennial crops such as Theobroma cacao, where experimental crosses are inherently time-consuming and logistically expensive. While LD has been evaluated in cacao, previous work has been focused on relatively narrow genetic bases. We use microsatellite marker data collected from a uniquely diverse sample of individuals broadly covering both wild and cultivated varieties to gauge the LD present in the different cacao diversity groups and populations. We find that genome-wide LD decays far more rapidly in the wild and primitive diversity groups of cacao as compared to those representing cultivated varieties. The impact that such differences can have on association analyses is demonstrated using phenotypic data on pod color and genotypic data from two cacao populations with contrasting patterns of LD decay. Our results indicate that the more rapid LD decay in wild and primitive germplasm can lead to higher-resolution mapping intervals when compared to results from cultivated germplasm. Through simulations, we demonstrate how future association mapping analyses, comprising of cacao samples with a wild or primitive background, will likely exhibit lower LD and would be more suitable for fine-scale association mapping analyses. As many traits targeted by cacao breeders are found exclusively in wild and primitive germplasm, association mapping in wild cacao populations holds significant promise for cacao improvement through marker-assisted breeding and emphasize the need to further explore the natural diversity of Amazonian cacao.