Surface topography plays a significant role in fabrication of different parts and has an effect on fatigue strength, corrosion resistance, creep life and surface friction. The quality of the in-contact surfaces for machined threads is also a significant factor in determining the clearance and durability of the joint. In this research the micro surface at the crest of the thread in high speed machining (HSM) is modelled using artificial neural networks. Feed rate, spindle speed and cutting fluid pressure were the inputs and surface roughness of the micro scale area (crest) of the thread was the output and the Taguchi L32 orthogonal array was used for design of experiment. Modelling and analysing the effective parameters were performed using multi-layer perceptron (MLP) artificial neural networks ANNs, which was shown to be highly adept for such tasks. To measure micro surface in the crest of the threads an accurate optical profile-meter with capability to measure surface roughness on the curvatures was used. Empirical tests were carried out to check the accuracy of the proposed model and verified that MLP-ANN is a strong and accurate application in modelling of micro surface in HSM. Moreover, the experimental outcomes show surface roughness is decreased by increasing cutting speed, decreasing the feed rate and setting the cutting fluid pressure in the range of 2-3.5 bars.