This paper proposes and discusses an active dual-sensor autofocusing method for measuring the positioning errors of arrays of small holes on complex curved surfaces. The dual-sensor unit combines an optical vision sensor and a tactile probe and is designed to achieve rapid automated measurements in a way that can be adapted to be suitable for deployment on a manufacturing machine tool. Mathematical analysis is performed to establish the magnitude of the deviation from the optimal focal length that is induced by the autofocussing method. This evaluation is based on the geometrical relationship and interaction between the radius of the tactile probe with both the measured holes and the complex-curved surface. A description is provided of a laboratory-based standalone dual-sensor autofocusing unit and test rig that was built to perform experimental validation of the method. This system is estimated to have a focusing uncertainty of 11 μm deriving mainly from the inaccuracy of the X-Z translation stage and the maximum permissible error of the tactile probe. A case study is presented which evaluates the accuracy of a pattern of ∅ 0.5 mm small holes on an elliptic cylinder. A mathematical analysis of that problem and practical results from both the tactile and optical sensors are provided and discussed. It is estimated that the deviation in optimal focusing induced by this automated method is between -23 μm and +95 μm. This is sufficiently accurate to ensure that the optical device can capture the entire space outline of each of the small holes on the complex curve surface clearly and can therefore identify its centroid from the image to provide a measurement of the position.