基于最小均方误差和准则的块式数据检测技术，是一种对时变色散信道数字信号进行有效检测的方法．该技术利用数据块中的已知训练符号对信道特性进行跟踪并对未知数据进行检测．采用简化的Ｃｈｏｌｅｓｋｙ矩阵分解算法，可减少系统复杂度和运算量．研究结果表明：该方案能有效地消除高频信道上的码间串扰，更适用于高频数字通信系统中的信号检测． Based on the minimum mean square error and criteria block data detection technology is a kind of time-varying dispersion channel digital signal for effective detection method. This technique uses the known training symbols in the data block to track the channel characteristics and to detect unknown data. Simplify the Cholesky matrix decomposition algorithm can reduce system complexity and computational complexity. The results show that this scheme can effectively eliminate the intersymbol crosstalk on high frequency channels and is more suitable for signal detection in high frequency digital communication system.