[Purposes] Large-scale sensor deployment and data redundancy are common issues in highway slope monitoring. Furthermore, existing sensor placement schemes generally lack a mechanism for dynamic adjustment of their effectiveness. Our paper proposes a placement optimization method that integrates Singular Value Decomposition (SVD) and a Greedy algorithm to address those problems. The aim is to achieve dynamic optimization of the monitoring layout and efficient reconstruction of the data. [Methods] First, observation data collected continuously over one month during the initial operation phase of an existing slope monitoring network were systematically organized. Subsequently, SVD was used to extract the dominant linear low-rank subspace from the observed signals. Within this subspace, the Greedy algorithm was combined to optimally select physical locations. These locations serve as candidate placement points. The selection criterion was the ability to effectively characterize the information from removed sensors with minimal reconstruction error. Finally, the complete high-dimensional state field was reconstructed within this low-dimensional subspace. The reconstruction was based on measured data from the selected positions. The least squares method was used for this reconstruction. To validate the effectiveness of the proposed method, a study was conducted using measured data. The data came from 24 GNSS sensors installed on a national highway slope in Zhejiang Province. The data covered one continuous month. [Findings] The results show that the proposed method can accurately reconstruct the displacement signals of the removed sensors. This reconstruction was achieved after removing 3 sensors. The accuracy level was a coefficient of determination greater than 0.9. [Conclusions] This demonstrates the superiority and reliability of the proposed method. This research provides a feasible technical approach for achieving dynamic optimization of slope monitoring networks in later stages, mutual characterization of monitoring data, and lightweight deployment.