Due to the intricacy characteristics of the Underwater Acoustic(UWA) channel especially the limited bandwidth, Orthogonal Frequency Division Multiplexing(OFDM) is used because of its high spectrum efficiency. However, relatively high Peak-to-Average-Power Ratio(PAPR) limits the efficiency of OFDM in UWA communication, leading to saturation in the power amplifier and consequent distortion of the signal. Clipping and C companding as the most classic and convenient algorithms, are widely applied to address the high PAPR issue. However clipping introduces additional noise which degrades the system’s performance and traditional C companding is also not suitable for underwater acoustic field. Thus, an improved C companding combined with clipping is proposed here. Due to the sparseness of clipping noise, Compressed Sensing(CS) can be utilized to estimate it. The scheme exploits pilot tones and data tones as observations instead of reserve tones, which is different from the previous works and improves data rate. Furthermore, to minimize the effect of the underwater acoustic channel, the channel is also estimated using the CS technique, which provides more accurate channel characteristics than Least Square(LS) or Minimum Mean Square Error(MMSE) estimation algorithms. The effectiveness of the proposed algorithm is proved in computer simulations as well as in a pool experiment. Due to the intricacy characteristics of the Underwater Acoustic (UWA) channel especially the limited bandwidth, Orthogonal Frequency Division Multiplexing (OFDM) is used because of its high spectrum efficiency. However, relatively high Peak- to-Average- Power Ratio (PAPR) limits the efficiency of OFDM in UWA communication, leading to saturation in the power amplifier and consequent distortion of the signal. Clipping and C companding as the most classic and convenient algorithms, are widely applied to address the high PAPR issue. However, clipping introduced additional noise which Thus, an improved C companding combined with clipping is proposed here. Due to the sparseness of clipping noise, Compressed Sensing (CS) can be utilized to estimate it. The scheme exploits pilot tones and data tones as observations instead of reserve tones, which is different from the previous works and improves data rate. Furthermore, to minimize the effect of the underwater acoustic channel, the channel is also estimated using the CS technique, which provides more accurate channel characteristics than Least Square (LS) or Minimum Mean Square Error (MMSE) estimation algorithms. effectiveness of the proposed algorithm is proved in computer simulations as well as in a pool experiment.