Abstract: Abstract: A compressed data gathering method based on compressed sensing (CS) for table grape cold-chain logistics was proposed, which aimed to not only monitor the quality and safety of the table grapes and improve the transparency, but also to gather the sensing data scientifically and efficiently during cold-chain logistics. CS is a new theory by which a signal can be recovered efficiently with just a few samples. The proposed method exploits the compressibility of the signal to reduce the number of samples required to recover the sampled signal at the remote receiver. All the sensor nodes will send their sensing data to the aggregation node when they receive the control instruction that was sent by the aggregation node at the initial time of the network. The sensor nodes will discard the abnormal data and acquire again, then they will go into sleep and wait for the next control instruction when the data is sent successfully. The aggregation node will send the compressed data that was measured by using the measurement matrix to the remote data receiver by the GPRS wireless technology. The remote data receiver will finish the reconstruction of the compressed data by using the reconstruction algorithm when the compressed data is received. We adopted the Gaussian random distribution matrix as the measurement matrix and the orthogonal matching pursuit algorithm as the data reconstruction algorithm, for they are the classical and efficient algorithm for the solution of the compressed sensing. In addition, we built the biorthogonal wavelet transform sparse matrix according to the characteristics of the sensing data to realize the sparse representation of the compressed data. Finally, we undertook the performance test of the method and the system under the simulation of cold-chain conditions located in the simulation lab of agricultural products traceability of College of Information and Electrical Engineering, China Agricultural University. The test included the reconstruction error of compressed sensing and the energy consumption of the nodes. The result showed that the method had a good compression result, whose data compression ratio can be 91%, and it could reconstruct the sensing data with high accuracy by transmitting the compressed data to the remote receiver. The absolute error of the reconstructed data of temperature and humidity was 0.07℃ and 0.05% respectively under the constant temperature, 0.15℃and 0.006% respectively under the variable temperature. The voltage decay rate of the aggregation node under the compression mode was less than that under the direct mode, which showed that the method we proposed could prolong the lifetime of the network efficiently.