A new method combining the pattern recognition (PR) technique with micro-PIXE spectrum was used for direct assessment of lead pollution in the atmosphere of Shanghai City. Single aerosol particles (SAP) of PM 10 (<10 µm) were analyzed using the nuclear microprobe. Every particle is characterized with its micro-PIXE spectrum, which can be considered its fingerprint. The PR technique was applied to trace a lead contaminated aerosol particle back to its source. The discrimination of different pollutant sources was enhanced with investigating the individual aerosol particles. The results showed that the lead contamination from automobile exhaust should not be neglected. The lead concentration with low level was detected in most unleaded gasoline particles; however, the highest lead level of 1500 ppm was found in one of them. Furthermore, four other main pollutant sources contributing to the lead contamination in the Shanghai atmosphere were clearly identified by this method. They are the cement industry, the coal combustion, the oil combustion, and the metallurgic industry. Some other unidentified particles suggested that some more lead emitters might also exist in Shanghai.