In recent years, wireless positioning technologies have been widely considered in the developments of new telecom systems and services. The increasing requirements of location-based services (LBS) in many practical applications have promoted the investigations of location estimation methods. With the system models and network structures addressed in the technical reports of 3GPP, the geometry-improved location estimation (GILE) algorithm is proposed to explore geometric relationshIPs between base stations (BSs) and mobile station (MS) in order to improve the accuracy and precision of location estimations. A geometric constraint based on the network topology and time-difference-of-arrival (TDOA) measurements is constructed to confine MS's location estimation within a specific closed region. Moreover, a geometric reformation derived from minimizing the geometric dilution of precision (GDOP) is utilized as an assistance to increase the degree of freedom in space domain against the errors from TDOA measurements to location estimations. According to the performance evaluations, proposed GILE algorithm is perceived to outperform other existing localization methods, especially under poor network topologies and insufficient measurement inputs.