Behavior of livestock in a pasture needs to be detected, and motion path and location information of livestock need to be collected in the intelligent pasture management, the acquisition of the above information can be realized by using the pasture Internet of things. The system architecture based on wireless sensor network can meet the communication requirements between different sampling nodes. However, because the location of a pasture is generally remote and network environment is complex, interruption and packet loss often appear in the connection with Internet. Using the traditional network connection mode can lead to a large number of data loss. In order to reduce the loss of sampled data in connection with Internet, a data transmission scheme based on opportunistic network is presented. Moreover, the communication status of sensors in a pasture environment is analyzed and three communication ways between sensors are summed up, which are modeled and analyzed respectively. Under the premise of the sensor limited storage capacity, the calculation method of accessing point density is put forward by using the principle of opportunistic network, and the relationship between livestock movement speed, sensor node storage capacity and data transmission loss rate is summarized, so as to ensure the system data loss in the design to allow for the range. Finally, an experimental evaluation and validation of the theoretical results of the method is given to prove the correctness of this program.