S hort-wavelength (15-150km) ocean phenomena are difficult to be observed by conventional altimeter due to the sampling method limited the resolution of the data. It is therefore the future wide-swath interferometric altimetry (WIA) missions like "S WOT" and "Guanlan" are designed to observe these phenomena. However, observing the short-wavelength dynamic process of the high eddy kinetic energy (EKE) region is still a challenge. The short-wavelength validation of WIA is also an urgent problem to be solved. In-situ observation platforms array can provide data set with high spatial and temporal resolution, which can capture short-wavelength dynamic process and validate the altimeter data. S imulation is a key to verify the feasibility of the observation strategy. In this paper, we carried out simulated observation experiments based on high spatial -temporal resolution numerical models and introduced motion model of mooring and keep-station glider to improve the accuracy of in-situ observation simulation under the influence of the dynamic ocean currents which optimized by deep learning methods. In this experiment, the steric height (S H), reconstructed by different observation array strategies and the sea surface height (S S H) of the measured positions are compared based on wavenumber spectrum. We tested the performance of the mooring array, station-keeping glider array, mooring and wave-energy profiling buoy (M&WPB) combination array and multi-device array. The results show that the glider and M&WPB arrays have better performance of strong current regions. It could be helpful for the observation of dynamic processes in high EKE regions and has potential application for the validation of WIA.