Accurate blood flow measurements during surgery can improve an operation's chance of success. We developed near-infrared spatio-temporal image spectroscopy (NIR-STICS), which has the potential to make blood flow measurements that are difficult to accomplish with existing methods. Specifically, we propose the technique and we show feasibility on phantom measurements. NIR-STICS has the potential of measuring the fluid velocity in small blood vessels (less than 1 mm in diameter) and of creating a map of blood flow rates over an area of approximately 1 cm(2). NIR-STICS employs near-infrared spectroscopy to probe inside blood vessel walls and spatiotemporal image correlation spectroscopy to directly-without the use of a model-extract fluid velocity from the fluctuations within an image. We present computer simulations and experiments on a phantom system that demonstrate the effectiveness of NIR-STICS.