To address the difficulties traditional slope reliability calculation methods face in considering the uncertainty among multiple variables and the large computational burden, this paper proposed a slope reliability analysis method based on the improved whale optimization algorithm (IWOA) and support vector machine (SVM). First, the basic theory of SVM was described. The differential mutation strategy and adaptive weight factor were introduced to improve the whale optimization algorithm (WOA), and the performance of the IWOA was tested. Then, the key parameters of SVM were optimized based on the IWOA, and the slope reliability analysis model was constructed. Finally, by taking a slope with an explicit performance function as Example 1, the reliability index of the slope was calculated based on IWOA-SVM. The results were compared with those of existing reliability methods, and the sensitivity of random variables was analyzed. By taking a general homogeneous slope without an explicit performance function as Example 2, the calculation results of IWOA-SVM, Monte Carlo simulation (MCS), and first order reliability method (FORM) were compared. The results show that the slope reliability analysis model based on IWOA-SVM has good fitting performance in both the global scope and the scope of checking points; especially within the scope of checking points, the fitting accuracy is higher. The slope reliability index obtained by IWOA-SVM is very close to that obtained by MCS, validating the accuracy of this method. IWOA-SVM is also applicable to slopes without explicit performance functions, validating the universality of this method. Compared with the MCS method, the IWOA-SVM method avoids a large number of samplings and significantly improves computational efficiency. Slope reliability is positively correlated with the internal friction angle φ and cohesion c and negatively correlated with the tensile crack depth z, the water-filled depth coefficient of tensile cracks i_w, and the horizontal seismic acceleration coefficient α. The random variable with the greatest influence on slope reliability is α, followed by i_w, c, and φ, while z has the least influence.