The multi-user Holographic Multiple-Input and Multiple-Output Surface (MU-HMIMOS) paradigm, which is capable of realizing large continuous apertures with minimal power consumption, has been recently considered as an energyefficient solution for future wireless networks, offering the increased flexibility in impacting electromagnetic wave propagation according to the desired communication, localization, and sensing objectives. The tractable channel modeling of MU-HMIMOS systems is one of the most critical challenges, mainly due to the coupling effect induced by the excessively large number of closely spaced patch antennas. In this paper, we focus on this challenge for downlink multi-user communications and model the electromagnetic channel in the wavenumber domain using the Fourier plane wave representation. Based on the proposed channel model, we devise the maximum-ratio transmission and Zero-Forcing (ZF) precoding schemes capitalizing on the sampled channel variance that depends on the number and spacing of the patch antennas in MU-HMIMOS, and present their analytical spectral efficiency performance. Moreover, we propose a low computational ZF precoding scheme leveraging Neumann series expansion to replace the matrix inversion, since it is practically impossible to perform direct matrix inversion when the number of patch antennas is extremely large. Our extensive simulation results showcase the impact of the number of patch antennas and their spacing on the spectral efficiency of the considered systems. It is shown that the more patch antennas and larger spacing results in improved performance due to the decreased correlation among the patches. In addition, it is demonstrated that our theoretical performance expressions approximate sufficiently well the simulated spectral efficiency, even for the highly correlated cases, thus verifying the effectiveness and robustness of the presented analytical framework.