The estimation of the force generated by an activated muscle is of high relevance not only in biomechanical studies but also more and more in clinical applications in which the information about the muscle forces supports the physician's decisions on diagnosis and treatment. The surface electromyographic signal (sEMG) reflects the degree of activation of skeletal muscles and certain that the sEMG is highly correlated to the muscle force. However, the largest disadvantage in predicting the muscle force from sEMG is the fact that the force generated by a muscle cannot be directly measured non-invasively. Indirect measurement of muscle force goes along with other unpredictable factors which influence the detected force but not necessarily the sEMG data. In addition, the sEMG is often difficult to interpret correctly. The sEMG-force relationship has been investigated for a long time and numerous papers are available. This review shows the limitations in predicting the muscle force from sEMG signals and gives some perspectives on how these limitations could be overcome, especially in clinical applications, by using novel ways of interpretation.