We propose a method based on the redshift dependence of the Alcock-Paczynski (AP) test to measure the expansion history of the Universe. It uses the isotropy of the galaxy density gradient field to constrain cosmological parameters. If the density parameter $\Omega_m$ or the dark energy equation of state $w$ are incorrectly chosen, the gradient field appears to be anisotropic with the degree of anisotropy varying with redshift. We use this effect to constrain the cosmological parameters governing the expansion history of the Universe. Although redshift-space distortions (RSD) induced by galaxy peculiar velocities also produce anisotropies in the gradient field, these effects are close to uniform in magnitude over a large range of redshift. This makes the redshift variation of the gradient field anisotropy relatively insensitive to the RSD. By testing the method on mock surveys drawn from the Horizon Run 3 cosmological N-body simulations, we demonstrate that the cosmological parameters can be estimated without bias. Our method is complementary to the baryon acoustic oscillation or topology methods as it depends on $D_AH$, the product of the angular diameter distance and the Hubble parameter.