Electrospinning is the process of extruding a fine fiber from a charged polymer solution. The fiber is continuously stretched by electrostatic forces and the evaporation of solvent while traveling through air prior to deposition. These fibers provide a high surface area to volume ratio and a high porosity which are useful in many engineering areas. Fiber diameter is one of the most important morphological properties of electrospun fiber as it is the main parameter for quality control. Small fiber diameter and higher fiber uniformity are desired in many applications. But one major issue with the process is the lack of a functional model that can link processing parameters and polymer solution properties to fiber morphology (fiber diameter and its distribution). In this study, adaptive neuro-fuzzy inference systems (ANFIS) and support vector machines (SVMs) models were used to establish a relationship between PEO nanofiber diameter and electrospinning processing parameters such as the polymer concentration, spinning surface distance, applied voltage and volume flow rate. The predictive performances of the two models were estimated and compared to those of multiple linear regression (MLR). The results indicated that the performance of SVMs was better than ANFIS and MLR methods. It was observed that the relationship existing between each electrospinning processing parameters and nanofiber diameter is nonlinear. The relative importance of each processing parameter was also computed.