A channel model is proposed to predict the time series of Ku band rain attenuation during a rain event at a tropical location. The model is based on considering the Gaussian distribution of the conditional occurrence of rain attenuation with a particular value of the attenuation occurring before. The mean (μ) and standard deviation (σ) for the distribution are modeled with the experimental data. The measured attenuation at a particular time instant is used to obtain μ and σ and to predict the attenuation after certain interval. The channel model has tested well giving the predicted attenuation that agrees with the measured value with a mean error within 15% higher than 1 dB. Validity of the model is also tested with the first-order and second-order statistics of attenuation occurrence, on a long-term basis. The method can also be applied even if attenuation measurements are missing for certain period of time with increased error. © 2011 John Wiley & Sons, Ltd.