A novel stitching interferometry method for the form measurement of large and moderately curved optical surfaces is presented. It is implemented on a measurement system for specimen dimensions of up to 1000 mm x 200 mm in area and with radii of curvature down to 2 m. The system has a 5-axis positioning system which can align the employed Fizeau interferometer to any surface forms within the specifications for subaperture measurements. The presented method is based on wavelength-swept interferometry measuring the topography and the cavity length between interferometer and specimen. External angle measurements with two tiltmeters and – depending on the test specimen – with an autocollimator or rotary encoders provide the relative position of specimen and interferometer. For the stitching, only the offset of the subapertures has to be fitted. Specimens ranging from flat to freeform surfaces are measured to study different aspects of the measurement setup. Among others, the instrument transfer function is determined. An analytic uncertainty budget according for the measured topographies is derived. Furthermore, a method is introduced to estimate the radius uncertainty for spherical specimens based on the topography uncertainty and a simple Monte Carlo simulation. A novel concept of a three-dimensional autocollimator-like rotation sensor with a measuring range of up to 10° is presented. It can be used in the stitching interferometry setup in the future. Finally, Monte Carlo simulations of the measurement process and the evaluation are performed for the measured specimen. Various error contributions of the system components are taken into account. The measurements, the respective uncertainty budgets and simulations show a good agreement. The stitching interferometry method is transferable to a measuring system for specimens of up to 1.5 m in diameter and ≥ 10 m in radius of curvature which is currently being set up. For this purpose, further Monte Carlo simulations were performed to find the uncertainty contributions with the highest sensitivity for different applications. This is the accuracy of the angle measurement, and for strongly curved specimens the accuracy of the lateral positioning. At the end of this thesis, possibilities for the further development of the methods used are outlined and discussed.