Traditional current sensors face several challenges, including large size, high cost, temperature sensitivity, low accuracy, and disruption of the original circuit structure. While optical and Hall sensors offer advantages such as non-contact measurement, their high cost, stringent environmental stability requirements, and complex maintenance procedures limit widespread adoption. The emergence of advanced magnetoresistance, such as Giant Magnetoresistance (GMR), Anisotropic Magnetoresistance (AMR), and Tunnel Magnetoresistance (TMR), has led to the development of non-contact current sensors. These sensors have demonstrated exceptional sensitivity and performance, particularly in power systems. This paper proposes an optimization method for calculating the transfer coefficient K in TMR-based linear array current sensors. The method maximizes signal-to-noise ratio (SNR), simplifying the iterative calculation process in AC systems and enhancing measurement accuracy. Simulation results confirm that the proposed method enables precise current measurements, even with uncertainties in relative positioning and sensitivity direction of the sensor. Additionally, it effectively mitigates noise interference. In the simulation, under DC conditions, current is calculated after 6 iterations, with error of 3.464 × 10−14A. Under AC conditions, initial iteration calculation of transfer coefficient K took 0.685 s, and current measurement maximum error is 1.769 × 10−6A.