Enclosures are commonly used to reduce the sound exposure of workers to the noise radiated by machinery. Some acoustic predictive tools ranging from simple analytical tools to sophisticated numerical deterministic models are available to estimate the enclosure acoustical performance. However, simple analytical models are usually valid in limited frequency ranges because of underlying assumptions whereas numerical models are commonly limited to low frequencies. This paper presents a general and simple model for predicting the acoustic performance of large free-standing enclosures which is capable of taking into account the complexity of the enclosure configuration and covering a large frequency range. It is based on the statistical energy analysis (SEA) framework. The sound field inside the enclosure is calculated using the method of image sources. Sound transmission across the various elements of the enclosure is considered in the SEA formalism. The model is evaluated by comparison with existing methods and experimental results. The effect of several parameters such as enclosure geometry, panel materials, presence of noise control treatments, location of the source inside the enclosure, and presence of an opening has been investigated. The comparisons between the model and the experimental results show a good agreement for most of the tested configurations.