Techniques have been developed to analyze walking gait using accelerometer and gyroscope data from miniature inertial measurement units (IMU), but few attempts have been made to use similar approaches for running gait. The purpose of this study was to develop an algorithm capable of estimating running speed using a single shank-mounted IMU. Raw acceleration and angular velocity were recorded from an IMU sensor attached on the lateral side of the shank in the sagittal plane and a method of reliably detecting the shank vertical and the minimal shank velocity gait event was used to segment a running sequence into individual strides. Through integration, the orientation of the shank segment was determined and an estimate of stride-by-stride running speed was calculated by integrating the acceleration data. The algorithm was verified using data collected from a group of seven volunteers running on a treadmill at speeds between 2.50 m/s and 3.50 m/s. Over the entire speed range, the estimation results gave a percentage root mean square error (%RMSE) of approximately 4.10%. With the accurate estimation capability and portability, the use of the proposed system in outdoor running gait analysis is promising.