Vanilla is an economically important crop for low-lying humid tropical regions. World demand for natural vanilla is increasing, but cultivated plants face serious phytosanitary problems. The disease known as Fusarium wilt is mainly related to the fungus Fusarium oxysporum f. sp. vanillae, and for its management, the pathogen–host relationship must be understood. Four in vitro multiplied vanilla accessions were evaluated: two Vanilla planifolia from Colombia and Mexico, one from V. odorata, and one (1) F1 hybrid (V. rivasii × V. trigonocarpa). In addition, three isolates of Fusarium from different symptomatic plants present in small-scale agroforestry systems: (1Fov) F. oxysporum f. sp. vanillae from leaf, (2Fov) F. oxysporum f. sp. vanillae from root and (3Fs) F. solani also from root. Plants with two months of growth were inoculated in vitro by immersion of roots, and the development of Fusarium wilt was recorded for 15 days, using a severity scale to describe symptoms and to calculate the area under the disease progress curve (AUDPC). No statistical differences were found when analyzing the interaction between Fusarium isolates and vanilla accessions. However, when independently analyzing the design factor Fusarium isolates, there were significant differences; the 1Fov isolate of F. oxysporum f. sp. vanillae induced the highest symptoms as well as death in some plants of all accessions, while F. solani was considered a secondary pathogen. There were no statistical differences for the vanilla accessions factor, but the values of AUDPC and symptoms observed suggest a slight resistance in all the accessions. Therefore, it is suggested to explore the vanilla gene pool to generate multiplication material with resistance genes and to contribute with genetic improvement to successfully integrate the management of Fusarium wilt in commercial systems.