The location information is very crucial for the sensing data in modern agriculture. However, positioning errors and sparse anchors are two key problems that should first be solved for the localization of the sensor nodes. We proposed a novel algorithm to tackle with these challenges. When the system of adjacent anchor distance equations is ill, a minimized-stress search algorithm (MSS) can decrease positioning error greatly. A collaborative sparse-anchored scheme (CSA) has an excellent positioning effect on low density of anchor, specifically on marginal sensor nodes. Our experimental result verified validity and accuracy of the algorithm. It improved feasibility and cost of WSN positioning technique, significantly.