Abstract. In the last period, the demand of prostheses has massively increased. To guarantee their reliability, properties of durability, biocompatibility, and osseointegration results to be mandatory. Possessing these attributes, Ti-6Al-4V alloy represents the most employed material for implants realization, and because of its microstructure, it can be manufactured by different processing methods, i.e., machining, and additive manufacturing. Considering the necessity of patient-tailored implants, and the capability of additive manufacturing to produce single batch, and complex shapes, at relatively low cost and short time, this latter represents a rewarding process. Compared to biocompatibility, that is mainly function of material chemistry, durability and osteointegration concern mostly surface roughness that affects cells growth at bone-prosthesis interface. After additive manufacturing process and prior to be inserted in the human body, a prosthetic implant is finished by machining operations, hence, the attainment of an appropriate resulting surface roughness is crucial for obtaining a successful implant. Thus, roughness forecasting capability, as a function of the employed finishing process, permits its optimization, avoiding expensive scraps. For this reason, this paper deals with the development of predictive models of surface roughness when micro-milling Ti-6Al-4V alloy specimens.