Low-frequency (LF) magnetic communication in the 30 Hz–3 kHz band is suitable for offshore observation and cross-medium information transmission because magnetic fields experience relatively low attenuation in conductive media such as seawater and soil. Previous studies have demonstrated the feasibility of high-temperature Superconducting QUantum Interference Device (HTc SQUID) receivers for LF magnetic communication. However, further quantitative investigation is required regarding transmitter-structure optimization, magneto-quasi-static near-field model validation, and coordinated transmitter–receiver design under unshielded geomagnetic conditions. In this study, an unshielded LF magnetic communication link is established, and a propagation model based on the magneto-quasi-static near-field approximation is developed using the experimental parameters. The applicability of this approximation is further verified through electrical-size analysis. The transmitter employs a symmetrically spaced double-loop antenna consisting of two 23-turn Litz-wire loops, each with a radius of 0.25 m and separated by 0.25 m, whereas the receiver is a high-sensitivity system based on an HTc SQUID gradiometer. Measurements are conducted at 91, 533, and 911 Hz over distances of up to 5.7 m using a 23-turn single-loop antenna, the double-loop antenna, and a 46-turn single-loop antenna. A comparison under the same total number of turns separates the gain produced by the double-loop configuration from that resulting solely from an increased number of turns. Off-axis measurements at axial distances of 1 and 2 m are also performed to quantify transmitter-receiver alignment tolerance. End-to-end communication performance is evaluated using a 96 Hz amplitude-modulated carrier, a data rate of 6 bit/s, and a 13-bit Barker sequence, with signal demodulation achieved through adaptive-threshold detection and sliding-code frame synchronization. The HTc SQUID gradiometer exhibits an equivalent magnetic-field-gradient noise of 71 pT·m^-1·Hz^-1/2 at 100 Hz and can stably detect weak magnetic signals in the presence of geomagnetic and power-line interference. At 91 Hz and a transmission distance of 2 m, the measured magnetic flux density is 21.16 nT with a relative standard deviation of 0.36%, whereas the simulated value is 22.84 nT, corresponding to a deviation of 7.36%. At 911 Hz, the double-loop antenna increases the received magnetic flux density by up to approximately one order of magnitude relative to the 23-turn single-loop antenna and extends the maximum reliable communication distance from 4.0 m to beyond 5.7 m, representing an improvement of more than 42%. Under the same total number of turns, the double-loop antenna still provides an average gain of approximately 3.25 dB over the 46-turn single-loop antenna, confirming that the enhancement originates from the antenna configuration rather than simply from the increased number of turns. When the off-axis angle approaches 45°, the gain decreases to approximately 1.3-1.4 dB, thereby defining a practical transmitter-receiver alignment boundary. In the communication experiment, 38,136 bits are successfully demodulated at a signal-to-noise ratio of 41.78 dB without any observed bit errors, yielding a 95% confidence upper bound of 7.87×10^-5 for the bit-error rate. Further tests show that the bit-error rate increases rapidly as the signal-to-noise ratio approaches the reliable communication threshold of approximately 15 dB. These results establish a complete model-experiment-communication validation framework and demonstrate that the coordinated design of the double-loop transmitter and high-sensitivity SQUID receiver can substantially improve LF magnetic communication performance in unshielded environments. The principal contributions of this study are the quantitative separation of antenna-configuration gain from turn-number gain, identification of the off-axis performance boundary, and statistically supported verification of end-to-end communication reliability. The results provide a basis for transmitter optimization, receiver configuration, and alignment control in future cross-medium magnetic communication systems.