Drug delivery by topical application has higher patient acceptance and lower morbidity than intraocular injection, but many ophthalmic treatments are unable to enter the eye or reach the posterior segment after topical application. The first stage towards posterior segment delivery after topical application is ocular surface penetration and existing models are in vivo or use large quantities of tissue. We therefore developed a novel ex vivo model using discs of porcine and human cornea and sclera (5 mm diameter) to assess penetration of a candidate neuroprotective siRNA. siRNA against caspase 2 or control solutions of known penetrance were applied to the corneal epithelial surface and trans-corneal penetration and corneal adsorbance measured at fixed time points. To demonstrate that leakage did not occur, we applied dextran blue, which should not penetrate the intact cornea and did not do so in our model. Fluorescein penetration (0.09%) was less than rhodamine B (6.98%) at 60 min. siCASP2 penetration was 0.01% by 60 min. When the applied siCASP2 was washed off after 2 min, (representing lacrimal drainage) 0.071% penetrated porcine cornea by 60 min and 0.0002% penetrated human cornea and 0.001% penetrated human sclera. Our ex vivo model rapidly and cost-effectively assesses transcorneal penetration of candidate topical therapies, allowing rates of trans-corneal penetration for potential therapies such as siRNA to be evaluated with small quantities of human or animal tissue.