While software testing is essential for enhancing a software system's quality, it can be time-consuming and costly during developing software. Automation of software testing can help solve this problem, streamlining time-consuming testing tasks. However, generating automated test data that maximally covers program branches is a complex optimization problem referred to as NP-complete and should be addressed appropriately. Although a variety of heuristic algorithms have already been suggested to create test suites with the greatest coverage, they have issues such as insufficient branch coverage, low rate of success in generating test data with high coverage, and unstable results. The main objective of the current chapter is to investigate and compare the coverage, success rate (SR), and stability of various heuristic algorithms in software structural test generation. To achieve this, the effectiveness of seven algorithms, genetic algorithm (GA), simulated annealing (SA), ant colony optimizer (ACO), particle swarm optimizer (PSO), artificial bee colony (ABC), shuffle frog leaping algorithm (SFLA), and imperialist competitive algorithm (ICA), are examined in automatically generating test data, and their performance is compared on the basis of various criteria. The experiment results demonstrate the superiority of the SFLA, ABC, and ICA to other examined algorithms. Overall, SFLA outperforms all other algorithms in coverage, SR, and stability. © 2024 Elsevier Inc. All rights reserved.