One of critical design requirements for wireless sensor networks (WSN) is the accurate acquisition of locations of their nodes. This paper presents the positioning methods in WSN based on the linear frequency modulation continuous wave (FMCW) techniques. Reference nodes (RNs) with known locations transmit linear FMCW reference signals, while other sensor nodes estimate their locations based on the time frequency difference arrival (TFDA) or time frequency of arrival (TFOA). The geometric locations of RNs affect the location accuracy. In order to overcome the geometric-dilution-of-precision effect and improve the location accuracy, the data fusion methods are proposed. The location estimates from TFDA and TFOA are combined by using Bayes rule and fuzzy logic methods. The simulation results show that the data fusion method enjoys higher location accuracy.