This paper develops a new recognition method utilizing the phase information to improve the performance of radar high resolution range profile (HRRP) target recognition system. The statistical properties of complex HRRPs are analyzed, and three common Gaussian models, which are adaptive Gaussian classifier (AGC) model, jointly Gaussian (JG) model and probabilistic principal component analysis (PPCA) model, are generalized to the complex domain to model complex HRRPs. It is demonstrated that the structures and parameter estimators of the three complex models are invariant to the initial phase. Furthermore, to enhance the recognition performance under low signal-to-noise ratio (SNR) conditions, a noise robust modification algorithm is introduced. Experimental results show that the proposed method can obtain higher average correct recognition rates. Also, the modified models can deal well with the noisy test samples.