Thin-walled parts are often used in various industrial applications and due to their functional requirements, higher accuracies are generally desired. But due to their low stiffness characteristics, deformation and chatter problems are frequently encountered in their machining, thus resulting in poor accuracy. Moreover, due to 5-axes milling, which is usually required for finish machining of such parts, further complexities in the process are added and, consequently, achieving higher accuracy becomes more challenging. Therefore, in this study, the influence of tool inclination angle and feeding direction has been investigated on the resultant surface accuracy of thin cantilever shaped parts in finish milling conditions through experiments, finite element method simulations and theoretical discussions. Moreover, since work material’s microstructure and hardness also have key influence on its machinability, therefore, the effect of heat treatment state on resulting surface accuracy also has been explored. Three different titanium alloy, Ti–6Al–4 V specimens have been used. Two of the specimens were given solution treatment at 1050 ℃/1 h followed by aging at 550 ℃/4 h. One of the specimens after being solution treated was allowed to cool in air while the other was cooled in water to obtain different microstructure and hardness whereas the third specimen was used in as-received condition (casted). The three Ti–6Al–4 V specimens were machined at five different tool inclination angles (60°, 70°, 75°, 80°, and 85°) with two feeding directions (horizontal inwards and vertical inwards). Results have shown that the angle near the perpendicular to the surface i.e. 85° (5° away from the perpendicular) has lowest deformation value and also has better surface quality, which shows that the lowest effective cutting speed at this angle has helped in achieving higher accuracy. Moreover, among the cutting directions studied, vertical inward direction produces smoother surface due to its less pushing effect on the surface being cut as compared to horizontal inwards direction. And in terms of heat treatment state, the specimen which was cooled in air after being solution treated has produced better results for all cutting conditions studied, mainly attributed to fine homogenous lamellar α + β Ti–6Al–4 V structure achieved due to slow cooling rate of air. Analysis of variance using response surface methodology has been carried out to develop predictive model and to study the influence of each variable on the accuracy. The developed model has shown good accuracy when validated through additional experiments.