Inkjet printed electrodes based on metal nanoparticle inks represent a significant component in low cost, thin film electronics. When scaled to continuous reel-to-reel processing platforms, there is an advantage in non-contact, imaging-based methods to monitor the quality of inkjet printed structures in real time. We developed a machine learning method to predict inkjet printed electrode sheet resistance based on microscope images of the device. The method can be extended to nondestructive, uninterrupted quality monitoring many reel-to-reel thin film manufacturing applications.<br/> Strips of silver nanoparticle electrodes are systematically printed with different inkjet print parameters such as ink drop size and drop spacing. Then, a machine learning model is trained on processed microscope images of the electrodes and experimentally measured electrode sheet resistance. The resulting model can predict sheet resistance from images of the electrode with error as small as 10%.