We consider low-noise angular velocity estimation for serial link manipulators using inertial readings from rate gyros and linear accelerometers. The research is founded on microelectromechanical systems (MEMS) components, which offer an attractive alternative to many traditional angular sensors due to their low cost, low power requirements, small size, and straightforward “strap-down” installation. By using a multi-MEMS configuration, an algebraic estimate of angular acceleration, where low- and high frequency perturbations are mostly proportional to the physical distances of linear accelerometers, is fused with rate gyro readings with the well-known principles of complementary and Kalman filtering. Experiments on a robotic three-link planar arm rig and a hydraulic heavy-duty manipulator demonstrate the feasibility of our practically lag-free novel approach.