Artificial muscles, especially in fiber format, are a key enabler for wearable and human assistive robotics. Practical use requires both high power density and good controllability. In this paper, we present closed-loop position control of a robotic arm actuated by agonistic Electrofluidic Fiber Muscles (EFMs): the ensemble of fiber ElectroHydroDynamic (EHD) pumps and thin McKibben actuators. One EFM drives a 70 g robotic forearm, whose elbow joint angle (range 5° to 75°) is controlled using a hybrid feedback–feedforward strategy that combines a proportional–integral–derivative (PID) controller with a quasi-static inverse model, identified offline from experimental data. We report successful position control with average rise time of 0.44 s, settling time of 2.44 s, and mean steady-state error of 0.2°. The hysteresis of EFMs is dramatically reduced. These results pave the way for practical use of Electrofluidic Fiber Muscles in wearable and human-centric robotic systems.