Abstract: Measurement errors of the relative distance from an accident well to its relief well were calculated from magnetic ranging tools with higher measurement accuracy and lower operational risks on the premise that the accident well was within the tools’ detection ranges. As the error calculated by the “trapezoidal” error model could not be coupled with the wellbore trajectory error, a measurement error model for magnetic ranging tools was built through the covariance propagation rate. The measurement errors of magnetic ranging tools were coupled with the wellbore trajectory errors to obtain the total covariance matrixes in the working planes of the magnetic ranging tools. The error ellipses and the optimal selection method of magnetic ranging tools for the relief well were thereby derived. The example calculation showed that when the current bottom of the relief well was 7.41 m away from the accident well, the semi-axis lengths of the measurement error ellipses were 1.26 m and 0.33 m. After the ranging operation, the relief well drilling continued to drill. The wellspot tool or RGR-I tool were recommended when the relative distance between the two wells was 6.68 m or was shortened to 5.21 m and then to 2.07 m, respectively. The research results show that the decreasing trend of the relative distance between the two wells is greater than the increasing trend of the measurement error of the relative distance after the range measuring operations and as relief well drilling continues. When the detection range of a magnetic ranging tool fully covers the measurement error ellipse of the relative distance between the two wells, magnetic ranging tools with shorter measurement ranges but higher accuracy can be used.