Small variations in the geometry of Micro Electro Mechanical Systems (MEMS) can yield very large variations in performance. Variations in geometry are a consequence of the MEMS fabrication process, and are unavoidable with current fabrication technology. To achieve quick, accurate, and precise measurements of MEMS geometry, we have previously reported on our pioneering use of capacitance to measure MEMS geometry. The ability to capacitively probe MEMS geometries has the potential to more precisely obtain geometric uncertainties, and to realize autonomous on-chip measurements in-the-field. The precision of measurement method depends on the precision of the capacitance meter, which is subject to various sources of noise. In this paper, we examine the effect of this noise using our off-chip capacitive measurement method. In our present approach, we consider four sources of noise and analyze how they individually contribute to the uncertainty in the extraction of MEMS geometry. The four sources of noise are: noise from the voltage source, internal noise of the capacitance meter, noise from external electromagnetic fields, and thermal noise. We verify our analytical results with simulation and validate our results with experiment. With off-chip capacitive probing, we find that the uncertainty in geometric extraction is most strongly affected by external electromagnetic fields, moderately affected by noise from the measurement equipment and thermal noise, and least affect by the applied voltage. We measure the uncertainty in geometry due to shielded and unshielded conditions, and we predict the uncertainty in geometry due to the voltage source and thermal noises.