We describe the implementation of main second-order gradient methods fur unconstrained nonlinear optimization in the programming language LISP. This approach ensures significant improvements of the known procedures in procedural programming languages,