The channel capacities of various lading channels are calculated and compared with that of Gaussian noise channel to find out the decrements of channel capacity according to each fading environment. As a result, it is confirmed that the channel capacities in Rician and m-distribution fading channels approach to that of Gaussian noise channel as direct-to-indirect power ratio in Rician fading channel and fading index m in m-distribution fading channel increases respectively. And the difference between two channel capacities of Gaussian noise channel and each fading channel which is dependent on carrier-to-noise power ratio (CNR) is found. Also the improvement of channel capacity of Rayleigh fading channel by introducing two-branch diversities is obtained. For diversity reception, predetection maximal-ratio and postdetection selective combining techniques are adopted. The results show that the improvement of channel capacity by predetection maximal-ratio combining diversity is superior to the postdetection selective combining diversitiy regardless of correlation coefficient between two diversity branches in Raylelgh fading channel. The best improvement is achieved when two branches are noncorrelative in both two diversify techniques and as correlation coefficient of two diversity branches is smaller, the improvement of channel capacity is greater.