Your question is ... ???

I have some experiences in Eviews coding.

What you need to do is to learn the Eviews object "logl" and express the 3*3 matrix in the log-likelihood function of BEKK. 

就看看看

brooks的书编的比较早，书中介绍的记得是3.0的版本。如果有较新的6.0的版本的话
1.先分别计算a,b,c时间序列的GARCH,在 proc选项卡内选择make garch covariance series，取得garch 序列
2.把a,b,c的garch series合并到一个system内，在proc内选择Estimate
3.在Estimate method中选择arch,下面的arch model spcification model type中选择diagonal bekk,就可以了

4# zhaopengyong
先分别计算a,b,c时间序列的GARCH,在 proc选项卡内选择make garch covariance series，取得garch 序列,
应该是make garch variance series选项，取得GARCH序列吧？