The article considers the problems of implementation and usage of mathematical packages in mathematical disciplines and also in the performance of course and degree works of students of mathematical and engineering professions. One of the important problems is the lack of methodological support packages. Here are viewed the most well-known and widely used mathematical packages such as MathCAD, MatLab, Mathematics and Maple in Windows and their application in teaching general course of mathematics, numerical methods of the multidimensional statistical analysis and other disciplines. For each package gives their capabilities, advantages and disadvantages. The recommendations in any disciplines use them more effectively. We give examples of execution of tasks of linear algebra students of the first course and laboratory works on numerical methods for students of the second course of Baumann Moscow State Technical University using MathCAD. For senior courses we consider packets intended for data processing and data analysis. The list of literature includes the books available in the library of the BMSTU.