This study evaluates the Radial Basis Function Partition of Unity Method (RBF-PUM) for itsrobustness and suitability for solving PDEs in option pricing. The robustness was tested byexamining the relative error behavior across its input parameters. To increase accuracy and toreduce computational time, three modifications were tested. These changes consisted ofchanging the node layout to use clustered Halton nodes, least squares projection aspreprocessing, and adding linear polynomials. The study shows that RBF-PUM achieves robustresults, suggesting its applicability to other option pricing problems. Furthermore, using a Haltonnode layout proved to be beneficial for the accuracy when using a smaller amount of nodes.The original implementation was less accurate, but more stable as the amount of nodesincreased. Least squares projections and added polynomials offered minor accuracy gains butincreased computation time, often making the trade-off unfavorable.