Comparison of different indices in evaluating the association between socio-economic (SE) inequalities and disease is important to select the most efficient indicators for intervention policies. Therefore, SE indices of different methodological inspiration were compared to analyse the effects of deprivation on cancer mortality by cause at local levels. Four indices, two drawn from the European Deprivation Index (EDI) for Italy and two from the Italian Deprivation Index (IDI), were computed for Umbria region and Genoa city by two different techniques (classification by quintiles vs. by normalization). A fifth one, the Socio-Health Deprivation Index (SHDI), was calculated and normalised specifically for the studied populations. ANOVA with F-test (p < 0.05) measured the statistical significance of cancer deaths by socio-economic status defined by the different methods. Colorectal, lung, prostate, breast and hematologic cancers were analysed by gender using the above indices. Results obtained from EDI and SHDI reproduced associations between SE and cancer mortality reported in scientific literature better than IDI (e.g., the linear negative trends emerging for breast and prostate). This is probably due to their computing procedure (bottom-up selection of the index variables). Most of differences derived from the clustering choice. In fact, the normalisation improved the performances of both the EDI and the IDI. Normalised EDI and the SHDI better identified the needs of populations in socio-health intervention fields. Nevertheless, for a deeper comprehension of association between deprivation and disease distribution by strata of age and gender, the contemporary use of all the indices is advisable.