At present, 19 double neutron star (DNS) systems are detected by radio timing and 2 merging DNS systems are detected by kilo-hertz gravitational waves. Because of selection effects, none of them has an orbital period $P_b$ in the range of a few tens of minutes. In this paper we consider a multimessenger strategy proposed by Kyutoku et al. (2019), jointly using the Laser Interferometer Space Antenna (LISA) and the Square Kilometre Array (SKA) to detect and study Galactic pulsar-neutron star (PSR-NS) systems with $P_b sim$ 10-100 min. We assume that we will detect PSR-NS systems by this strategy. We use standard pulsar timing software to simulate times of arrival of pulse signals from these binary pulsars. We obtain the precision of timing parameters of short-orbital-period PSR-NS systems whose orbital period $P_b in (8,120),$min. We use the simulated uncertainty of the orbital decay, $dot{P}_{b}$, to predict future tests for a variety of alternative theories of gravity. We show quantitatively that highly relativistic PSR-NS systems will significantly improve the constraint on parameters of specific gravity theories in the strong field regime. We also investigate the orbital periastron advance caused by the Lense-Thirring effect in a PSR-NS system with $P_b = 8,$min, and show that the Lense-Thirring effect will be detectable to a good precision.