Microfluidic biochips integrate different biochemical analysis functionalities on-chip and offer several advantages over the conventional biochemical laboratories. In this thesis, we focus on the flow-based biochips. The basic building block of such a chip is a valve which can be fabricated at very high densities, e.g., 1 million valves per cm2. By combining these valves, more complex units such as mixers, switches, multiplexers can be built up and the technology is therefore referred to as microfluidic Very Large Scale Integration (mVLSI). The manufacturing technology for the mVLSI biochips has advanced faster than Moore’s law. However, the design methodologies are still manual and bottom-up. Designers use drawing tools, e.g., AutoCAD, to manually design the chip. In order to run the experiments, applications are manually mapped onto the valves of the chips (analogous to exposure of gate-level details in electronic integrated circuits). Since mVLSI chips can easily have thousands of valves, the manual process can be very time-consuming, error-prone and result in inefficient designs and mappings. We propose, for the first time to our knowledge, a top-down modeling and synthesis methodology for the mVLSI biochips. We propose a modeling frame-work for the components and the biochip architecture. Using these models, we present an architectural synthesis methodology (covering steps from the schematic design to the physical synthesis), generating an application-specific mVLSI biochip. We also propose a framework for mapping the biochemical applications onto the mVLSI biochips, binding and scheduling the operations and performing fluid routing. A control synthesis framework for determining the exact valve activation sequence required to execute the application is also proposed. In order to reduce the macro-assembly around the chip and enhance chip scalability, we propose an approach for the biochip pin count minimization. We also propose a throughput maximization scheme for the cell culture mVLSI biochips, saving time and reducing costs. We have extensively evaluated the proposed approaches using real-life case studies and synthetic benchmarks. The proposed framework is expected to facilitate programmability and automation, enabling the emergence of a large biochip market.