This work proposes SeamFarm, a novel framework for distributed data analytic for precision agriculture. SeamFarm is proposed based on the seamless computing paradigm. In general, nodes deployed for precision agriculture consists of different types of devices. We choose seamless computing since it provides the capabilities that fog computing does not have, such as dealing with the heterogeneity of devices. Also, the high number of devices resulted in an extensive amount of data that should be analyzed, whereas most of IoT devices are resource-constrained devices. Thus, we consider distributing not only the data but also the processing task to all of the available nodes. The results show that SeamFarm can maintain the CPU usage of IoT devices. Also, SeamFarm can meet all of the functional and non-functional requirements of distributed data analytic for precision agriculture.