This thesis describes the fabrication of plastic microneedle devices, their fabrication by injection molding, and analysis of the penetration mechanics. Injection molding is an economical mass-production technique that may encourage widespread adoption of microneedles for drug delivery. Four polymers were injection molded into hexagonal and square patterns of between 91 and 100 needles per array. The patterns and geometries were chosen to study the effect of needle spacing and array design on penetration force. Two needle spacings of approximately 1 mm and 1.5 mm were employed for both patterns. Molded parts showed tip radii below 15 microns, heights of 600 to 750 microns, and an included angle of approximately 30 degrees. An economic analysis performed of the injection molded polymer devices showed that they can be manufactured for approximately $0.10 - $0.179 per part, which should be low enough to gain market acceptance. The added benefits of low pain perception, improved drug delivery for certain treatments, and the possibly of being recyclable make injection molded micro-needle devices a desirable alternative to silicon or metal microneedles. Penetration tests were performed with plastic micro-needle arrays and arrays of steel needles of the same spacings and patterns. Silicone rubber with mechanical properties similar to human skin was used as a skin simulant. The results showed that the micro-needles penetrated skin to depths between 120 and 185 microns depending on pattern, spacing, tip radius and needle length. This depth is sufficient to deliver drug therapies, but not so far that they stimulate the nerve endings present beyond 130 microns inside the dermis layer in human skin. An analytical model was developed to estimate the effects of various microneedle and skin characteristics on penetration force. The model was based on literature sources and derived from test results. The model accounted for coefficient of friction, tip radius, tip angle, and needle spacing, as well as the skin mimic's mechanical properties such as elastic modulus, mode I fracture toughness, and puncture fracture toughness. A Monte Carlo simulation technique was used to correct for errors in needle length and testing angle. Comparison of the experiments to the model showed good agreement.