We propose a novel modeling and control framework for the autonomous excavator with main control valve (MCV), which distributes fluid from pumps to hydraulic actuators with the number of the pumps less than that of the actuators and whose internal hydraulic circuitry switches depending on operating conditions and internal pressures. We first derive the mathematical model of the MCV, including the switching components and supply pump flow constraint. We then design a novel velocity-field control for the bucket position/orientation, which, by relying on a constrained-optimization formulation, can adjust the velocity-field following speed reflecting the physical constraints imposed by the MCV in such a way that the bucket fully follows the desired velocity-field when the constraints are inactive or still preserves the desired direction (or automatic stopping) while slowing down when the constraints become active (e.g. flow saturation). We further show that this optimization can be reduced to simple real-time solvable formulation with its solution existence/optimality (or suboptimality) guaranteed. Simulation is also performed to verify the theory by using a detailed Simulink/Sim-Hydraulics model. (C) 2019 Elsevier Ltd. All rights reserved.