This paper presents a joint synchronization and localization algorithm based on the time-difference-of-arrival (TDOA) for asynchronous Wireless Sensor Networks (WSNs), where the positions of anchors are relatively fixed but unknown. To obtain time synchronization and anchor positions, each anchor broadcasts signals periodically, while other anchors receive the broadcasting signals and stamp times-of-arrival (TOAs). Based on these TOAs, we estimate the internal clock parameters and anchor positions in three steps: least square estimation (LSE) of the relative clock skew based on TDOAs, maximum likelihood localization (MLE) of anchors using biased time of flight (TOF), and LSE estimation of relative clock offsets. Anchor pairs stamp the TDOAs when a tag (a wireless sensor node that requires localization) transmits a signal. The biased TDOAs, due to asynchronous local clocks, are compensated by relative clock offset estimation. The maximum likelihood estimation is used for the tag localization based on the Gaussian noise model. We evaluate the performance of the proposed synchronization and localization algorithm using the Cramer-Rao lower bound (CRLB). Simulations are carried out to verify the validity of the algorithm in this paper.