MEMS technologies promises to revolutionize health care by providing precise control of biological fluids for both diagnoses and treatments. For example, microneedles can be used for sample collection for biological analysis, delivery of cell or cellular extract based vaccines, and sample handling providing interconnection between the microscopic and macroscopic world. Microneedles may be used for low flow rate continuous drug delivery such as the continuous delivery of insulin to a diabetic patient. Microneedles are interesting from a design perspective not only because of their small size but because they provide a range of geometries and flow characteristics. This paper uses microneedle design as an example of the potential interaction between experiment and computation for the improved design of microfabricated microfluidic devices in general. Previously, fluid flow in microneedles was studied experimentally. Here, we use computational modeling capabilities in concert with experimental results to optimize the design of medical microneedles and, thus, to shorten the whole design/fabrication cycle. We compare CFD simulations to analysis and experiments for flows in three microneedle geometries - straight, bent and filtered (Fig. 1). The bent microneedle was found to have the highest fluid carrying capacity of 0.082 ml/sec at 138 kPa with a Reynolds number of 738. A microneedle with a built in microfilter 192 μm wide, 110μm high and 7mm long also had flow rates of 0.07 ml/sec (Fig. 2). Although the throughput of these microneedles is low they still compare favorably with other microneedle designs. Laminar flow models were found to accurately predict the flow behavior through the microneedles. All computational modeling was performed with the CFDRC CFD-ACE+ suite of software tools.