The invention discloses an elevator health degree detection method and device based on a statistical model. The elevator health degree detection method comprises the steps that firstly sample data affecting the elevator health degree is collected; then KMO testing and Bartlett testing are conducted on the sample data, and whether factor analysis is suitable for being conducted on the data or not is judged; variable communality testing is conducted, factor analysis is conducted by selecting common factors, and expressions of the common factors are written out; multiple linear regression analysis is conducted on the common factors, and an elevator health degree detection model is determined; and finally test data affecting the elevator health degree is collected, and the elevator health degree is obtained through the determined detection model. According to the elevator health degree detection method and device, the elevator health degree model is established by adopting a factor analytic method in the statistics, embedded software and hardware development is conducted on the basis, and the defect that an original elevator detection device only detects but not analyzes is overcome.