This paper proposes a hybrid control strategy of a novel linear piezoelectric walking stage based on two sorts of piezoelectric actuators, which takes the load variation into account. The proposed stage consists of two parallel 4-bar lever amplification mechanisms with flexure hinges actuated by piezoelectric stacks to heighten the vertical distance (that is more tolerable to the assembly discrepancy), two compression springs (that is able to maintain a fixed linear position without powering), and two shear piezoelectric actuators (that can achieve longer and equivalent to walking motion) in a small form factor. The proposed stage has two operating modes, namely a coarse positioning mode with a more extensive travel range and a fine positioning mode with a nanometer-level resolution, to possess excellent performance for the linear piezoelectric walking stage of load variations. One multimodal switching controller and one feedforward-feedback controller conduct the coarse mode and fine mode, respectively. The optimal frequency for a specific load is obtained through a backpropagation neural network in the multimodal switching control. In the feedforward-feedback control, the inverse mathematical model based on the Bouc-Wen hysteresis model is used to mitigate the hysteresis effect in the feedforward part while the proportional–integral–derivative controller in the feedback part handles the external system disturbances. Experimental results show the proposed hybrid coarse/fine mode control strategy's effectiveness to satisfy an efficient and accurate positioning task.