To support the extremely high spectral efficiency and energy efficiency requirements, and emerging applications of future wireless communications, holographic multiple-input multiple-output (H-MIMO) technology is envisioned as one of the most promising enablers. It can potentially bring extra degrees-of-freedom for communications and signal processing, including spatial multiplexing in line-of-sight (LoS) channels and electromagnetic (EM) field processing performed using specialized devices, to attain the fundamental limits of wireless communications. In this context, EM-domain channel modeling is critical to harvest the benefits offered by H-MIMO. Existing EM-domain channel models are built based on the tensor Green function, which require prior knowledge of the global position and/or the relative distances and directions of the transmit/receive antenna elements. Such knowledge may be difficult to acquire in real-world applications due to extensive measurements needed for obtaining this data. To overcome this limitation, we propose a transmit-receive parameter separable channel model methodology in which the EM-domain (or holographic) channel can be simply acquired from the distance/direction measured between the center-points between the transmit and receive surfaces, and the local positions between the transmit and receive elements, thus avoiding extensive global parameter measurements. Analysis and numerical results showcase the effectiveness of the proposed channel modeling approach in approximating the H-MIMO channel, and achieving the theoretical channel capacity.