This thesis examines user interaction with instrumentation in the specific context of space science. It gathers together existing practice in machine interfaces with a look at potential future usage and recommends a new approach to space science projects with the intention of maximising their science return. It first takes a historical perspective on user interfaces and ways of defining and measuring the science return of a space instrument. Choices of research methodology are considered. Implementation details such as the concepts of usability, mental models, affordance and presentation of information are described, and examples of existing interfaces in space science are given. A set of parameters for use in analysing and synthesizing a user interface is derived by using a set of case studies of diverse failures and from previous work. A general space science user analysis is made by looking at typical practice, and an interview plus persona technique is used to group users with interface designs. An examination is made of designs in the field of astronomical instrumentation interfaces, showing the evolution of current concepts and including ideas capable of sustaining progress in the future. The parameters developed earlier are then tested against several established interfaces in the space science context to give a degree of confidence in their use. The concept of a simulator that is used to guide the development of an instrument over the whole lifecycle is described, and the idea is proposed that better instrumentation would result from more efficient use of the resources available. The previous ideas in this thesis are then brought together to describe a proposed new approach to a typical development programme, with an emphasis on user interaction. The conclusion shows that there is significant room for improvement in the science return from space instrumentation by attention to the user interface.