1. A method for determining a precoding matrix indicator (PMI), comprising:receiving, by a user equipment, a reference signal sent by a base station; and

determining, by the user equipment, the PMI according to the reference signal, wherein the PMI corresponds to a precoding matrix W, and the precoding matrix W satisfies a second condition or a third condition;

sending, by the user equipment, the PMI to the base station; wherein

the second condition is that the precoding matrix W comprises one or more column vectors of a block diagonal matrix W1, or the precoding matrix W is obtained by performing weighted combination on one or more column vectors of a block diagonal matrix W1, wherein W1=diag{X1, . . . , XNB}, and NB?1, wherein at least one block matrix X is a product X=DV of a matrix D and a matrix V, and X?{X1, X2, . . . , XNB}; the matrix D is a diagonal matrix, D=?·diag{u1, u2, . . . , un, un*, un-1*, . . . , u1*}, ? is a complex factor, a complex number ui* is a conjugate complex number of a complex number ui, i=1, . . . , n, and n is determined by a quantity of antenna ports; and the matrix V is a constant modulus matrix; and

the third condition is that the precoding matrix W comprises one or more column vectors of a block diagonal matrix W1, or the precoding matrix W is obtained by performing weighted combination on one or more column vectors of a block diagonal matrix W1, wherein W1=diag{X1, . . . , XNB}, and NB?1, wherein at least one block matrix X is a Kronecker product of a matrix A and a matrix B, X=A?B, and X?{X1, X2, . . . , XNB}; the matrix A or the matrix B is a product of a matrix D and a matrix V; the matrix D is a diagonal matrix, D=?·diag{u1, u2, . . . , un, un*, un-1*, . . . u1*}, ? is a complex factor, a complex number ui* is a conjugate complex number of a complex number ui, i=1, . . . , n, and n is a quantity of rows of the matrix A or the matrix B; and the matrix V is a constant modulus matrix;

wherein in the second condition or the third condition, the precoding matrix W satisfies W=W1W2, wherein the matrix W2 is used to select one or more column vectors of the matrix W1; or is used to perform weighted combination on one or more column vectors of the W1 to obtain the precoding matrix W.