The RF receive coil array has become increasingly vital in current MR imaging practice due to its extended spatial coverage, maintained high SNR, and improved capability of accelerating data acquisition. The performance of a coil array is intrinsically determined by the current patterns generated in coil elements as well as by the induced electromagnetic fields inside the object. Investigations of the ultimate performance constrained by a specific coil space, which defines all possible current patterns flowing within, offer the opportunity to evaluate coil-space parameters (i.e., coverage, coil-to-object distance, layer thickness, and coil element type) without the necessity of considering the realistic coil element geometry, coil elements layout, and number of receive channels in modeling. In this paper, to mimic 7-T monkey RF head coil design, seven hypothetical ultimate coil arrays with different coil-space configurations were mounted over a numerical macaque head model; by using Huygens's surface approximation method, the influences of coil-space design parameters were systematically investigated through evaluating the spatial constrained ultimate intrinsic SNR and ultimate g-factor. Moreover, simulations were also conducted by using four coil arrays with limited number of loop-only elements, in order to explore to what extent the ultimate coil performance can be achieved by using practical coil designs, and hence several guidelines in RF coil design for monkey brain imaging at 7 T have been tentatively concluded. It is believed that the present analysis will offer important implications in novel receive array design for monkey brain MR imaging at ultra-high field.