Estimates of Earth’s gravity temporal variations by GRACE(−FO) have catalyzed a wide range of scientific studies and discoveries. Although an increase in the satellite pairs would reduce the error and increase the temporal and spatial resolution, mission costs limit populating additional GRACE-like pairs. One viable solution is to reduce costs by miniaturizing the satellite. As a first step in reaching this objective, the Miniaturized Prototype for GRavity field Assessment using Distributed Earth-orbiting assets (uPGRADE) project aims to produce a CHAMP-like prototype gravimetry satellite that includes star trackers, GNSS and accelerometers in CubeSat size. As one of the primary payloads, the utility of high-precision Micro-Electro-Mechanical Systems (MEMS) accelerometer for gravimetric mission has not been considered. Here, we evaluated three, six and nine MEMS arrangements. We found that the six MEMS parallel arrangement can observe both the desired non-gravitational accelerations and additional absolute value of the angular velocity. We developed a measurement error model, associated with MEMS position and orientation errors, to guide the MEMS optimal design and assembly. Finally, we conducted uPGRADE mission simulations using appropriate observations and model errors. The impact of a 10 nm/s2MEMS accelerometer error on gravity recovery is very close to that of the 5 mm GNSS error. However, the accelerometer error degrades the low-degree coefficients more significantly, particularly (Formula presented) and (Formula presented). The simulations indicate that the temporal gravity can be estimated up to degree 15, albeit with some compromise in the low-degree coefficients. Recommendations are made to lower the projected noise floor of MEMS accelerometer to enhance the low-degree coefficients accuracy.