A virtual space-time diversity soft direct-adaptation turbo euqalization (SDATEQ) is proposed for single-input-single-output (SISO) underwater acoustic (UWA) communications based on soft fractionally-spaced decision-feedback equalizer (SFSDFE) and virtual time-reversal mirror (VTRM). The SFSDFE combined with recursive least-squares (RLS) algorithm is employed to achieve virtual space diversity and increases the mutual information (MI) of iterations for faster convergence. Bidirectional structure with VTRM is exploited to mitigate error propagation phenomenon of DFE. An interpolator and digital phase locked loop (DPLL) are embedded into the euqalizer to overcome the residual Doppler frequency shift and recover the timing distortion. Results of simulations and field lake trial show that the proposed algothrim achieves better performance than Soft-DATEQ under the same equalizer order, and the bit error rate (BER) of 1500-m lake trial decreases from 1.50×10 -2 to 2.22×10 -5 .