For direct imaging of faint exoplanets, coronagraphs are widely used to suppress light and achieve a high contrast. Wavefront correction algorithms based on adaptive optics are introduced simultaneously to mitigate aberrations in the optical system. Stroke minimization is one of the primary control algorithms used for high-contrast wavefront control. This technique calculates the minimum deformation across the deformable mirrors' surface under the constraint that a targeted average contrast level in the search areas, namely the dark holes, is achieved. In this paper we present a modified linear constraint stroke minimization algorithm. Instead of using a single constraint on intensity averaged over all pixels, we constrain the electric field's real and imaginary part of each pixel in the dark holes. The new control algorithm can be written into a linear programming problem. Model reduction methods, including pixel binning and singular value decomposition (SVD), are further employed to avoid over-constraining the problem and to speed up computation. In numerical simulation, we find that the revised algorithm leads to more uniform dark holes and faster convergence.