Aiming at the problem of poor positioning accuracy of intelligent equipment, such as weak satellite navigation signal and phase locking caused by tree occlusion in Panax notoginseng planting area, a global navigation satellite system（GNSS）positioning algorithm for the shading environment of Panax notoginseng combine harvester based on pseudorange and Doppler double-difference positioning algorithm was proposed. Firstly, based on the difference of the influencing factors of pseudorange measurement and Doppler frequency shift measurement, the pseudorange double difference and Doppler frequency shift double difference were taken as inputs, and the carrier irritability ratio was used as the weight, and the measured values were fused through Kalman filter, so as to reduce the estimation error and correct the pseudorange and Doppler frequency shift measurement. Secondly, the Bayesian information criterion was used to select the regularization parameters, and the reweighted least squares problem was solved by Lasso regression to achieve the sparsity of the model and obtain the improved positioning results. Finally, the u-blox ZED-F9P high- precision GNSS receiver was used to collect the messages in RINEX format. Under three working conditions：open environment, shade shelter and tree shade shielding, the positioning accuracy test of the real vehicle was carried out. Compared with the traditional pseudo-distance positioning algorithm, open environment, shade shelter environment, and tree shade shade environment the position error was reduced by 13.43%, 56.08% and 46.35%, respectively, and the root mean square error of the positioning deviation was reduced by 75.64%, 62.31% and 50.21%, respectively. Under dynamic conditions, the positioning error was reduced by 36.97%, 52.14% and 62.37%, respectively, and the root mean square error of positioning deviation was reduced by 45.34%, 60.24% and 65.81%, respectively. The proposed method effectively reduced the positioning error caused by GNSS satellite signal difference and phase lock, and effectively improved the positioning accuracy and positioning credibility, The research result can provide theoretical and technical support for the problem of poor positioning accuracy of intelligent equipment due to tree shading in hilly and mountainous areas.