Incremental roughness prediction is a critical component of decision making of any pavement management systems, therefore, proper estimation is of paramount importance. This paper presents the application of functional equations and networks to incremental roughness prediction of flexible pavement. In the functional networks, neuron functions are multivariate and multiargumentative. Functional equations form the basis of functional networks, therefore, established theorem in functional equations are easily applicable in the analysis. The model is developed from validated set of incremental and interactive pavement distress functions in the highway design and maintenance standard models (HDM). The models proposed and developed are intended for use in infrastructure management (pavement) applications and as a performance model for pavement design. The paper presents a computational procedure of the functional network using the serial associative functional network. The functional equation and networks approach use the domain knowledge and data for developing the roughness models.Key words: pavement roughness, functional equation, functional networks, neural networks, pavement management.