Two kinds of the empirical orthogonal function analysis method for decomposing the mean-field are developed and a comparison is made between them. It is proved that the mean-field can be decomposed using eigenvectors obtained from anomaly-field decomposition, and that more information can be obtained. An example of sea bottom mean temperature analysis shows its remarkable effect in depicting the distribution features of various factors, such as cold water mass, currents and radiation.A few problems concerning the efficiency of the method are discussed and two matrices of relationship which represent the space-time characteristics of the field are derived. The formulae of space-time transformation are obtained conveniently. Two kinds of the empirical orthogonal function analysis method for decomposing the mean-field are developed and a comparison is made between them. It is proved that the mean-field can be decomposed using eigenvectors obtained from anomaly-field decomposition, and that more information can be obtained. An example of sea bottom mean temperature analysis shows its remarkable effect in depicting the distribution features of various factors, such as cold water mass, currents and radiation. A few problems concerning the efficiency of the method are discussed and two matrices of relationship which represent the space-time characteristics of the field are derived. The formula of space-time transformation are obtained conveniently.