Berezinskii-Kosterlitz-Thouless (BKT) transition, the topological phase transition to a quasilong range order in a two-dimensional (2D) system, is a hallmark of low-dimensional topological physics. The recent emergence of non-Hermitian physics, particularly parity-time (PT ) symmetry, raises a natural question about the fate of low-dimensional orders (e.g., BKT transition) in the presence of complex energy spectrum. Here we investigate the BKT phase transition in a 2D degenerate Fermi gas with an imaginary Zeeman field obeying PT -symmetry. Despite complex energy spectrum, PT -symmetry guarantees that the superfluid density and many other quantities are real. Surprisingly, the imaginary Zeeman field enhances the superfluid density, yielding higher BKT transition temperature than that in Hermitian systems. In the weak interaction region, the transition temperature can be much larger than that in the strong interaction limit. Our work showcases a surprising interplay between low-dimensional topological defects and non-Hermitian effects, paving the way for studying non-Hermitian low-dimensional phase transitions.Berezinskii-Kosterlitz-Thouless (BKT) transition, first discovered in the two-dimensional (2D) XY spin model [1-4], is a cornerstone for studying lowdimensional condensed matter physics [5]. In 2D systems, while Mermin-Wagner theorem forbids the emergence of a true long range order due to thermal fluctuations, BKT physics permits a topological phase transition to a quasi-long range order of topological defects (i.e., vortices) at very low temperature. Specifically, across a critical temperature T BKT , free vortices bind together spontaneously and form bound vortex-antivortex (V-AV) pairs, giving rising to a quasi-long range order. BKT transitions have been experimentally studied in a wide range of systems, including liquid helium films [6], 2D magnets [7-10], superconducting thin films [11][12][13][14], and 2D atomic hydrogen [15], where only macroscopic properties of the systems can be measured. In recent years, ultracold atomic superfluids have emerged as a versatile playground for the studies of BKT physics [16][17][18][19][20][21][22][23][24][25][26][27][28][29] with access to the underlying microscopic phenomena such as the visualization of the proliferation of free vortices. Experimental signatures of BKT physics have been observed in harmonically trapped , and the recent realization of ultracold superfluids in box potentials [30][31][32][33][34] offers a uniform platform for exploring BKT transitions.