Torsion dystonia is characterized by sustained muscle contractions causing twisting and repetitive movements and abnormal postures. The diagnosis can be made difficult, delayed, and often misled by several factors: variability of dystonia presentation, uncertain recognition of the specific physical signs, lack of diagnostic tests, wide etiological spectrum, and coexistence of other movement disorders. Diagnostic tools are of limited assistance for the diagnosis of dystonia, which remains based on clinical diagnostic skills. We propose here, a new diagnostic algorithm to systematize the clinical diagnostic workout. A correct recognition of the physical signs that constitute the hallmark of most dystonia syndromes provides the grounds to perform a structured diagnostic sequence and share a consistent methodology. This clinical algorithm may be enhanced by adding diagnostic tools for dystonia, once their diagnostic value is assessed.