This paper presents a semi-robotic method that combines force-controlled adaptive drill feed rate and human-robot collaborative strategies to enhance performance of non-repetititve micro-drilling tasks. Different materials afford different drill feed rates for optimal performance. To support the non-repetitive nature of the drilling task where the workpiece material may change from one task instance to the next, the drill feed rate is automatically adapted such that the reaction force sensed by the robot matches a threshold force setting, which is set as a function of the maximum axial-load capacity of the micro-drill. This ensures that the feed-rate and the interaction force gradually build up from zero, resulting in a better interaction between the drill bit and the surface. Aspects of drilling destabilization and non-repetitivity with respect to drill locations are addressed by applying human-robot collaboration in two contexts: 1) Human provides a complementary stabilizing support to the robot by holding its end-effector during task operation, which results in better drill quality. This approach is inspired by humans’ capability to stabilize unstable dynamics while performing contact-based tasks by using selective control of arm mechanical impedance that generates resistance forces against any destabilization caused during drilling and thereby enhancing drilling performance. 2) A human kinesthetically teaches a set of drill coordinates by physically holding the robot and guiding it to the designated locations, which avoids excessive programming effort needed for path planning, saving both time and computational cost. Experimental validation of the proposed approach was carried out by performing micro-drilling of diameters up to a minimum of 0.5 mm. The potential of the proposed approach was illustrated in the domains of composite repairs and bone-drilling in orthopedic surgeries.