With the rapid development of smart grid and IoT technology, low-power and high-precision current monitoring technology has become a key requirement for the intelligent upgrade of power systems. Magnetoelectric composite materials have the advantages of simple structure, convenient preparation, and high magneto electric conversion coefficient, and their detection limit for magnetic fields can reach the pT level. This article focuses on the magnetic electric coupling characteristics and sensing applications of magneto electric composite materials. Firstly, a hybrid model based on principal component analysis and random forest (PCA-RF) is proposed to accurately predict the magneto electric coefficients through dimensionality reduction and nonlinear regression. The overfitting problem of traditional methods in multi parameter coupling scenarios is solved, and the significant influence of magneto electric composite ratio and bias magnetic field on performance is revealed; Secondly, this article proposes a tag based differential sensor, which is designed with dual sensitive units based on magneto electric composite materials. Combined with U-shaped elastic fixtures, it can adapt to wires of different diameters and has significant resistance to common mode noise; On this basis, this article utilizes the energy conversion characteristics of magneto electric composite materials to achieve self powering, combined with ultra-low power signal processing circuits, and optimizes the static power consumption to approach zero through DC bias, further reducing the power consumption of current monitoring. Experimental verification shows that the low-power current monitoring method proposed in this paper has excellent performance in both magneto electric coefficient and low-power performance verification.