<p>Separate waste collection is one of the key tasks of waste management. Improving economic efficiency and reducing the negative environmental impact of transportation is crucial for collection companies. This article presents the results of efficiency improvements in terms of reducing collection time, shortening collection vehicle routes, and reducing exhaust emissions. In addition, an evaluation of currently performed collections of separately collected waste based on real-life collections in rural areas and towns in southern Poland is included. Artificial intelligence algorithms—tabu search and ant colony algorithm—were used to optimize routes. Route optimization resulted in significant reductions in collection lead times, reducing collection times for currently completed routes from 4.3 to 13.1% and the route length from 17.2 to 31.3%. Optimization resulted in an average 22% reduction in NO_x, PM, and CO₂ emissions. The results are beneficial to waste collection companies that do not use artificial intelligence to improve routes and collection plans. Reduced collection duration following optimization may facilitate the alternating schedule of waste collections and include more collection locations every shift. The findings demonstrate economic advantages for waste collection companies and beneficial environmental effects for residents in the waste collecting zone, including reduced emissions. Additional multi-criteria analysis can assist in choosing the optimal options for route design, vehicle selection, and possible fleet replacement with electric trucks or alternative fuel vehicles. </p>