Speeding　related-crashes　have　caused　numerous　fatalities　and　become　a　worldwide　health　problem.　This　study　aims　at　investigating　the　factors　affecting　injury　severity　in　speeding-related　crashes,　considering　the　spatial　heterogeneity　on　rural　and　urban　roads.　The　data　on　speeding-related　crashes　were　extracted　from　the　Crash　Report　Sampling　System　(CRSS)　between　2018　to　2020,　including　information　about　the　characteristics　of　drivers,　vehicles,　crashes,　roads,　and　the　environment.　Two　separate　correlated　random　parameter　order　probit　models　with　heterogeneity　in　means　(CRPOPHM)　were　established　for　speeding-related　crashes　on　rural　and　urban　roads,　and　the　plausibility　of　separately　modelling　injury　severity　was　tested　by　a　set　of　LR　tests.　The　model　results　showed　that　some　factors　were　significant　in　both　models,　while　others　were　significant　in　only　one　particular　model.　For　example,　heavy　trucks　and　weekends　are　significant　in　the　rural　model;　and　young　drivers,　rear-end　crashes,　speed　limits,　and　nights　with　lit　roads　are　significant　in　the　urban　model.　The　results　of　correlations　and　heterogeneity　in　means　of　random　parameters　of　the　two　models　also　showed　some　similarities　and　differences　for　speeding-related　crashes　on　rural　and　urban　roads.　Based　on　these　results,　some　policy　recommendations　are　proposed　to　mitigate　the　injury　severity　of　speeding-related　crashes.