Recently, the field of Wireless Sensor Network (WSN) has attracted a lot of attention from the scientific and research community due to its abilities in disaster prevention, weather forecasting, healthcare, and military applications. Indeed, the information brought by a sensor is still incomplete without knowing exactly where the event occurred. Hence, localization plays a vital role in designing a WSN. In this paper, we tackle one issue that resides in Distance Vector Hop (DVHOP) algorithm, more precisely the weak accuracy of DVHOP. So mainly, the reason for the mentioned problem is still the least squares method used by this algorithm. At this point, many solutions have been proposed to increase the accuracy of DVHOP such as DVHOP based on RSSI calculation, a combined DVHOP with meta-heuristic based nature, adding genetic steps to DVHOP positioning process, and so on. In our contribution, we have changed the least squares method of DVHOP by using the Fmincon solver that exists in MATLAB. Experimental results prove that the improved DVHOP iDVHOP outperforms the traditional DVHOP by handling the number of nodes and the percentage of anchors.