We report on molecular-dynamics simulations of the drag force experienced by a smooth sphere as it approaches a smooth planar surface to test the predictions of classical hydrodynamic theory. We use a simple repulsive Lennard-Jones-like model to represent the fluid interactions, and calculate the total force on the sphere as a function of its radius, velocity, and distance from the surface. We find that the presence of static solvation forces complicates the testing of hydrodynamic theory which predicts a divergent repulsive lubrication force as the gap vanishes. The solvation force contribution is most prominent at small gaps and small velocities. For a smooth wall its presence can lead to a total force that is oscillating between positive and negative, quite different from the hydrodynamic prediction. To enable an improved test of the lubrication predictions, we propose a different approach that measures the total force for approaching as well as receding spheres. We suggest a simple general analysis that decouples the dynamic and static force contributions on the sphere. The new decoupling method is applicable to simulations and laboratory experiments alike. We illustrate its power by applying it to the molecular-dynamics data.