Pulmonary embolism (PE) is a common, lethal yet treatable disease. The clinical diagnosis of PE remains to be a problem due to the nonspecific presenting signs, symptoms, electrocardiographic findings, arterial blood gas abnormalities and chest X-ray changes. Despite these nonspecific clinical findings, clinicians are adept at assigning pretest probability using overall clinical assessment. Clinical models have been developed to improve the accuracy of pretest probability assessment. D-dimers are becoming a widely available clinical tool useful in the diagnostic management of suspected PE. The limitations of the imaging modalities for PE [ventilation-perfusion (V/Q) scanning, spiral computerised tomography, pulmonary angiography and venous leg imaging] necessitate the use of these tests in series and in combination with clinical pretest probability assessment and D-dimer in diagnostic management algorithms. These algorithms permit safe diagnostic management of patients with suspected PE while limiting invasiveness, inaccessibility and expense.